CN1864088A - Fast scanner with rotatable mirror and image processing system - Google Patents

Abstract

A scanner for obtaining an image of an object placed on an at least partially transparent platform, wherein the platform has at least a first scan area and a second scan area. The scanner includes a white area formed at least partially around the edge portions of the platform with a plurality of markers, one rotatable mirror, one or more image sensors. In scanning, each partial image from each of the scan areas includes an image of at least one portion of the plurality of markers. An image processing system uses the image of the at least one portion of the plurality of markers in each of the consecutive partial images as a reference to combine the consecutive partial images so as to form a substantially complete image of the object corresponding to a full scan of the first scan area and the second scan area.

Description

Fast scanner with rotatable mirror and image processing processing system

This application is a PCT international patent application filed on October 7, 2004 by Ruiling Optics, an American company that is an applicant in all countries except the United States, and Yuping Yang, a US resident designated only as an applicant in the United States.

Several references, including patents, patent applications, and various publications, are cited in the Reference List and discussed in the present specification. Citation and/or discussion of these references are provided solely to enable clarity of the description of the present invention and are not an admission that these references are "prior art" to the invention described herein. All references cited and discussed in this specification are hereby incorporated by reference in their entirety with the same effect as if each reference were cited individually.

technical field

The present invention relates to a scanning device, and more particularly to the use of an image sensor and a rotatable mirror for high-speed image scanning and the use of marker images for high-speed image processing.

Background technique

Electronic storage of documents is becoming a common practice due to many significant advantages, such as ease of information sharing and management, saving physical space, and less risk of data loss. Once images of documents are stored in a computer system, a number of techniques exist, primarily in the form of software, to properly label, index, store, package, and retrieve these images. Therefore, the demand for scanning paper documents into electronic files or documents has increased dramatically in recent years. There are currently two scanning devices on the market that can be used to convert paper documents into electronic documents. The first is the so-called glass-surface flatbed scanner, which has a scanning speed of 0.033 to 0.143 pages per second. The second is a sheet-fed scanner with a scanning speed of 3 pages per second. Sheet-fed scanners can efficiently scan documents that have a uniform physical shape. Saving scanning time becomes especially important when the number of document pages to be scanned is large. Since not all documents can be entered through the input chute of said sheet-fed scanner, a flatbed device is essential for both office and personal use. For example, in hospitals, outpatient clinics, and accounts receivable departments of various companies, a large number of checks and payment instruction pages are received in the mail every day. Although these paper documents are good candidates for electronic storage, because they vary widely in size and shape, and are often folded and bound together in some fashion, scanning these documents into computer systems is lacking due to speed constraints. Valid devices in . Until today these paper documents could only be sorted, tagged, packaged and retrieved manually.

A current commercial flatbed scanner consists of a scanning head with a light source, a mirror, a focusing lens, and an optical sensor. All parts in the scan head move together during scanning. The sensor receives an optical signal from a document and converts the optical signal into an electrical signal, which is then processed into an image of the document.

Among all factors, the factors that limit the scanning speed of flatbed devices are the moving speed of the scanning head, the linear scanning rate of the optical sensor, the data transmission speed and the image processing speed. Image processing speeds that greatly exceed the speed of movement of the scanning head can be achieved. Data transfer speed depends on the selected protocol. The most common Universal Serial Bus (hereinafter "USB") port has a transfer speed of 1.5MBps (megabytes per second). It only takes 0.2 seconds to transfer a 300KB image file. Others such as Small Computer System Interface (hereinafter referred to as "SCSI") are several orders of magnitude faster than USB ports. A new generation of high-sensitivity optical sensors can scan up to 46,000 lines per second, for example, DALSA IT-P1-2048 (DALSA Corp., Waterloo, Ontario, Canada). If each page has 4,000 lines, the sensor can scan in less than 0.1 second. The fact that current sheet-fed scanners and copiers can scan 190 pages per minute (ppm) has proven that image processing speed, data transfer speed, and sensor linear rate are not speed bottlenecks.

The speed bottleneck of the platform device is that the scanning head moves slowly. Specifically, it is not that the stepper motor cannot drive the scan head fast enough, but rather the back and forth and start-stop motion of the scan head limits the speed of the scan head. Therefore, although both microprocessor speed and memory density have increased by several orders of magnitude in recent years, the increase in the scanning speed of the flatbed scanner has only been a gradual process.

Compared to the 0.033 to 0.143 pages per second scan time required by glass-surface flatbed scanners, cameras using area sensors can capture document images in an instant. An example of such an image scanning system is disclosed in US Patent No. 5,511,148, entitled "Interactive Copying System." Another embodiment of an image scanning system is disclosed in US Patent No. 6,493,469. It should be understood that U.S. Patent No. 6,747,764 also discloses a "camera box" similar device in which an area sensor faces up to capture an image of a document facing down. Said document is placed on a transparent platform. However, cameras with area sensors often do not have sufficient resolution to replace common office scanners. Similar to glass surface scanning devices, the flashing light emitted to the scanning area during scanning can cause discomfort and injury to the user, despite the high height of the device due to the need to maintain sufficient distance between the document to be scanned and the area sensor The device therefore has a limited scanning area.

Cameras with line sensors, known as line scan cameras, can produce higher image resolution than cameras with area sensors. However, scanning documents with a line scan camera has many inconveniences, for example, it is usually necessary to place the document to be scanned face up. Otherwise, in order to place the line scan camera under the document to be scanned, it is necessary to construct a bulky scanning device. In addition, this line scan camera device requires strong scanning light, which causes discomfort to human eyes during frequent image capture.

Based on the principle of a line scan camera, a scanner with a rotating mirror can be used as a glass surface (or flat-bed) scanner. But there are many problems to be solved before rotating mirror scanners become ubiquitous products for everyday use. A key issue is the required distance between the in-line sensor and the scanning area. Figures 1 and 2 show a scanner using a line sensor and a rotating mirror. Place the document to be scanned face down on the surface of scanning area 1. The rotating mirror 2 reflects the imaged light of the original document to the condenser lens 3 , and then sends it to the line sensor 4 . The rotating mirror rotates about axis 5 . This type of image scanner can scan original documents at very high speeds. However, this type of scanner is bulky and presents several other problems. As shown in Figure 2, the viewing angle _α0 is defined as the angle between the image path of the original document and the surface plane 1 of the original document located at the far end of the scanning area. Even after scanning image distortion is eliminated, the smaller the viewing angle α ₀ , the lower the extracted resolution is at the original document located near the far edge of the scanned area. In order to maintain a certain level of resolution, the viewing angle α ₀ must be greater than a certain threshold at all times during the scan. Therefore, in order to scan an original of size _L0 , the height _H0 of the scanner cannot be too small.

Many attempts have been made to shorten the distance between the document and the rotatable mirror and condenser lens. For example, US Patent No. 6,396,648 employs a fisheye lens and US Patent No. 6,324,014 employs a set of lenses. However, the reduction in distance achieved by using different lenses is limited and further complicates the overall system as a side effect of increased distortion of the captured image.

US Patent No. 6,493,469 acquires two partial document images with two area sensor cameras. Each partial image has relatively little distortion and good resolution. The two images are merged to form a complete image of the original document. This design has several problems. Image capture devices such as digital cameras employing expensive area sensors often do not have sufficient resolution to replace common office scanners. The design also requires documents to be placed face-up and the camera down, taking up a relatively large footprint, so it's not as convenient for frequent scanning as a flatbed scanner. The proposed method of combining two partial images is based on captured images, which is not reliable in achieving a high-quality combined image.

US Patent No. 5,909,521 also discloses a method of obtaining a complete document image from multiple partial images. But image processing is complex, and the quality of the alignment of images changes as parts of the scan are scanned. Therefore, this method does not provide a fast and reliable way to combine partial images into a whole.

Another problem in designing a rotating mirror scanner is precisely timing the coordination between image processing and the angular coordinates of the rotating mirror. Various methods have been proposed, for example, U.S. Patent No. 6,088,167 discloses a method of using a dedicated light sensor in addition to an image capture optical sensor for regularly capturing a beam of light when it is at a certain position to achieve scanning position Measurement.

US Patent No. 5,757,518 proposes several methods of timing the rotation of a rotating mirror. The first method is to count the main scan period, where no additional hardware for scan timing is required. But timing the main scan cycle is difficult to implement if multiple partial images of the same original document need to be obtained before being combined into a complete image. Also, cyclic calculation errors may accumulate. The second method is to measure the angular displacement of the mirror. An obvious disadvantage of this method is the need for additional components. For the same reason, the third method using an optical path length detector is not ideal.

US Patent No. 5,253,085 uses a synchronous sensor to detect the angular coordinates of a rotating mirror. In US Patent No. 5,973,798 additional sensors and hardware devices are used to detect angular coordinates.

A third problem in rotating mirror scanner design is the correction of non-uniform shading of the image due to non-uniform exposure. It is known in the scanner field that a standard white reference is required for obtaining the reference luminosity in order to find or calculate shading data. Many methods have been developed for shading correction, such as US Patent No. 6,061,102, US Patent No. 5,724,456, US Patent No. 6,546,197, US Patent No. 6,195,469, US Patent No. 5,457,547, and US Patent Publication No. U.S. 2003/0,142,367. However, in rotating mirror scanner designs, the non-uniformity of shading across the scan area is greater than in standard flatbed scanners. Hence the need for a larger "standard white reference" area.

A fourth problem in image scanner design is to remove distortions in the raw image originally acquired by the sensor. The captured raw image needs to be processed to obtain an undistorted image of the scanned document. There are many methods for removing distortion, such as US Patent No. 6,233,014, US Patent No. 5,253,085, and US Patent No. 6,219,446.

Yet another problem is that strong light shines outside the scanning area during scanning. For a quick scan, one approach is to use a bright light to cover the entire scan area. This can be very uncomfortable for the operator's eyes. Another method is to use a strong and narrow beam to scan the scan area below the image scan line. The second method minimizes the amount of illuminating light emitted outside the scanning area during scanning. However, the light beam needs to scan the scanning area completely synchronously with the image scanning. A solution enabling the second method has not been provided so far.

Accordingly, there remains a hitherto unaddressed need in the art that can address the above-mentioned disadvantages and deficiencies.

Contents of the invention

In one aspect, the invention relates to a scanner for obtaining an image of an object located on an at least partially transparent platform. The at least partially transparent platform has a first scan area and a second scan area, and each of the first scan area and the second scan area of the at least partially transparent platform has a first edge and a second edge, respectively.

In one embodiment, the scanner has a light source adapted to emit light. The scanner also has a rotatable mirror adapted to receive light from a first direction and reflect said light in a second direction to scan a partial image of an object located on the at least partially transparent platform, the rotatable mirror The mirror is further adapted to receive the scanned partial image of the object from a third direction opposite the second direction and to reflect the scanned partial image of the object in a fourth direction opposite the first direction. Furthermore, the scanner has a fixed mirror located on the optical path between the rotatable mirror and the first scanning area of the at least partially transparent platform, the fixed mirror is adapted to receive the light from the rotatable mirror in the second direction. reflecting light and reflecting said light received from the rotatable mirror towards a first scanning region of the at least partially transparent platform to scan a partial image of the target, and said fixed mirror also receiving the scanned partial image of the target, The partial image of the scanned target is then reflected toward the rotatable mirror in a third direction. In addition, the scanner has an image sensor for receiving a scanned partial image of the object from the fourth direction and outputting an electronic signal corresponding to the received scanned partial image of the object. Furthermore, the scanner has an image processing system for receiving electronic signals from the image sensor and recording the electronic signals in digital format. The scanner also has a condenser lens located on the optical path between the rotatable mirror and the image sensor, and a rotation device for rotating the rotatable mirror.

The rotatable mirror and the fixed mirror are arranged such that when the rotatable mirror is rotated, the rotatable mirror changes the second direction of light so that in only one full rotation of the rotatable mirror, reflection from the fixed The corresponding light reflected by the mirror along the fifth direction is from the first edge to the second edge of the first scanning area along the first scanning direction A and from the first edge to the second edge of the second scanning area along the first scanning direction B in sequence Sequential partial images of the scanned object. The image processing system combines the partial images recorded here to form a substantially complete target image, which is equivalent to full scanning along the first scanning direction A and the second scanning direction B respectively. In one embodiment, the at least partially transparent platform, the rotatable mirror, and the fixed mirror are arranged such that a first angle _α1 is defined as the at least partially transparent platform and the first angle connecting the lower edge of the fixed mirror and the first scanning area. The angle between the optical paths of the edges, and the second angle α ₂ is defined as the angle between the at least partially transparent platform and the optical path connecting the intersection of the first direction and the second direction and the first edge of the second scanning area, Both angles _α1 and _α2 are greater than a predetermined threshold angle α.

The rotatable mirror includes a flat mirror having at least one reflective surface. In one embodiment, the rotatable mirror has a polygon mirror. In one embodiment, the fixed mirror includes a plane mirror. In another embodiment, the fixed mirror comprises a curved mirror. In one embodiment, the image sensor includes at least one of a line sensor, an area sensor, and a combination thereof. In one embodiment, the light source includes at least one of a laser, a fluorescent tube, a light emitting diode assembly, a tungsten lamp, a tungsten halogen lamp, a halogen lamp, a xenon lamp and any combination thereof. The at least partially transparent platform includes a flat plate of at least partially transparent material. In one embodiment, the at least partially transparent platform comprises a glass plate or a transparent plastic plate.

In another aspect, the invention relates to a scanner for obtaining an image of an object located on an at least partially transparent platform, wherein the at least partially transparent platform has at least a first scanning area and a second scanning area, and the first scanning area and the Each of the second scan areas has a first edge and a second edge, respectively.

In one embodiment, the scanner has at least one light source for emitting light. Furthermore, the scanner has at least one rotatable mirror for receiving light from a first direction and reflecting the light in a second direction to scan an object on the at least partially transparent platform The partial image of the scanned target is received from a third direction and the scanned partial image of the target is reflected toward a fourth direction. Also, the scanner has at least one image sensor for receiving a partial image of the scanned object from a fourth direction and outputting an electronic signal corresponding to the received partial image of the scanned object. Furthermore, the scanner has an image processing system for receiving electronic signals from at least one image sensor and recording the electronic signals in digital format. The scanner also has at least one condensing lens located on the optical path between the at least one rotatable mirror and the at least one image sensor, and a rotating device for rotating the at least one rotatable mirror.

The at least one light source, at least one rotatable mirror and at least one image sensor are arranged such that the first direction and the fourth direction define a first angle (180°-β), and the second direction and the first direction Three directions define a second angle (180°+β), wherein the value of β is in the range of -15° to 15°, and when the at least one rotatable reflector is rotated, the at least one rotatable reflector changes the second angle of light Two directions such that, preferably in only one full rotation of said at least one rotatable mirror, said light travels along a first scan direction A from a first edge to a second edge of the first scan area and along a first scan direction B Continuous partial images of the target are sequentially scanned from the first edge to the second edge of the second scanning area. The image processing system combines the partial images recorded here to form a substantially complete object image equivalent to a full scan along the first scan direction A and the second scan direction B, respectively.

In yet another aspect, the invention relates to a method for obtaining an image of an object located on an at least partially transparent platform. Wherein the at least partially transparent platform is defined by a plurality of edge portions and the at least partially transparent platform has at least a first scan area and a second scan area.

In one embodiment, the method further includes forming a white area with a plurality of marks at least partially surrounding an edge portion of the at least partially transparent platform, each mark being disposed at a predetermined position of the white area. Each of the plurality of marks is identifiable from the white area. In one embodiment, at least a part of the white area is suitable for standard reference white. The method further comprises the step of sequentially scanning successive partial images of the target from the first scanning area and the second scanning area, respectively, wherein each successive partial image includes an image of at least one of the plurality of markers. Furthermore, the method includes using an image located on at least one of the plurality of markers in each successive partial image as a reference to combine the successive partial images to form a substantially complete full scan corresponding to the first scan area and the second scan area. target image. The adopting step further includes a step of correcting the formed target image, and trimming the image of the white area and the images of the plurality of marks from the corrected target image to obtain the target image, respectively.

In yet another aspect, the invention relates to a scanner for obtaining an image of an object on an at least partially transparent platform, wherein the at least partially transparent platform is defined by a plurality of edge portions and has at least a first scanning area and a second scanning area. scan area.

In one embodiment, the scanner has a plurality of marked white areas formed on at least part of the surrounding edge portion of the at least partially transparent platform, wherein each mark is arranged at a predetermined position in the white area, and a plurality of Each mark is recognizable from the white area. In one embodiment, at least a part of the white area is suitable for standard reference white. Also the scanner has optical means for sequentially scanning successive partial images of the object from the first scanning area and the second scanning area, respectively, wherein each successive partial image comprises an image of at least one of the plurality of markings. In one embodiment, the optical device includes at least one image sensor. The at least one image sensor includes one of a line sensor, an area sensor and combinations thereof. The scanner also has an image processing system for combining successive partial images using an image of at least one of the plurality of markers located in each successive partial image as a reference to form a region corresponding to the first scan region and the second scan region. A full scan of the area for a substantially complete image of the target. In one embodiment, the image processing system has a controller that performs correction by trimming an image of a white area and images of a plurality of markers from a corrected target image to obtain the target image, respectively. Steps for forming the target image.

In yet another aspect, the invention relates to a method for obtaining an image of an object located on an at least partially transparent platform, wherein the at least partially transparent platform has a plurality of scanning areas. In one embodiment, the method includes sequentially scanning successive partial images of objects from respective ones of the plurality of scan regions and combining the successive partial images to form a base corresponding to a full scan of the plurality of scan regions. Steps for complete target image.

In one embodiment, the at least partially transparent platform also has a plurality of marks, each mark being arranged at a predetermined position. Each successive partial image includes an image of at least one of the plurality of markers. The step of combining includes using an image of at least one of the plurality of markers in each successive partial image as a reference.

In another aspect, the invention relates to a scanner for obtaining an image of an object on an at least partially transparent platform, wherein the at least partially transparent platform has a plurality of scanning areas. In one embodiment, the scanner has optical means for sequentially scanning successive partial images of objects from respective ones of the plurality of scanning areas, and for combining the successive partial images received from the optical means Processing means thereby forming a substantially complete image of an object corresponding to a full scan of a plurality of scan areas.

In one embodiment, the optical device includes at least one image sensor. The at least one image sensor includes at least one of a line sensor, an area sensor and a combination thereof. The at least partially transparent platform has a plurality of marks respectively disposed at predetermined positions. each successive partial image includes an image of at least one of a plurality of markers used as a fiducial,

These and other aspects of the invention will become more apparent from the following description of a preferred embodiment, taken in conjunction with the accompanying drawings, where changes and modifications may be made without departing from the spirit and scope of the novel concepts disclosed.

Description of drawings

Figure 1 shows a perspective view of a conventional rotatable mirror-based scanner;

Figure 2 is a schematic side view of the scanner shown in Figure 1;

Figure 3 is a schematic perspective view of a scanner according to an embodiment of the present invention;

Figure 4 shows a schematic side view of the scanner shown in Figure 3, with an optical illustration of the scanner;

FIG. 5 is a block diagram showing an image processing flow of a scanner according to an embodiment of the present invention;

Figure 6 is a schematic bottom view of an at least partially transparent platform having a white area with a plurality of indicia in accordance with an embodiment of the present invention;

7 is a schematic diagram of a process for identifying scanline locations using markers according to an embodiment of the invention: (a) a portion of an at least partially transparent platform with markers, and (b) a magnified image of markers with scanlines at different locations;

Figure 8 shows the electronic signals corresponding to the images of the marks in Figure 7: (a)-(d) correspond to the electronic signals of the images marked in different positions respectively;

FIG. 9 is a schematic diagram showing the pre-alignment of two partial images of an image using a common marker according to an embodiment of the present invention;

Fig. 10 shows a schematic flow chart of combining two partial images scanned by a scanner according to an embodiment of the present invention: (a) two partial images scanned along the first scanning direction A and the second scanning direction B respectively, (b ) a pre-alignment of the partial images, (c) a combination of these partial images, and (d) a processed image of the partial images;

FIG. 11 is a schematic flow chart of the image shading intensity correction process;

Fig. 12 shows (a) a schematic perspective view of a scanner according to an embodiment of the present invention, and (b) the three-dimensional shading intensity distribution of two partial images scanned by the scanner;

Figure 13 is a schematic diagram of the shading intensity correction process;

Figure 14 is a cross-sectional view of the shading intensity distribution in Figure 13 extracted along the x-axis: (a) before the shading intensity correction process, and (b) after the shading intensity correction process;

Figure 15 is a cross-sectional view of the shading intensity distribution in Figure 13 extracted along the y-axis;

Figure 16 is a schematic diagram of the image distortion removal process using a marked image according to an embodiment of the present invention: (a) and (b) show how to generate a coordinate grid, (c) and (d) show how the distorted image The iterative process of the upper fine coordinate grid, (e) and (f) show how to perform image distortion removal correction on the pixel coordinates located at the P position in the coordinate grid, (g) in the distorted image that has not yet undergone distortion removal non-uniform distance between scan lines, and (h) showing the process of matching the stored coordinates of the marker and grid with the actual image of the marker and grid;

17 is a schematic diagram of a scanner according to an embodiment of the present invention: (a) a side view of the scanner, and (b) the optical geometry of the scanner;

FIG. 18 is a schematic diagram of a scanner according to an embodiment of the present invention;

Figure 19 is a schematic side view of a scanner with a light shield according to an embodiment of the present invention;

Fig. 20 is a schematic side view of a scanner according to an embodiment of the present invention;

Fig. 21 is a schematic side view of a scanner according to another embodiment of the present invention;

Fig. 22 is a schematic side view of a scanner according to an alternative embodiment of the present invention;

Figure 23 shows the geometry for calculating the curvature of the curved auxiliary mirror in Figure 22;

FIG. 24 is a schematic diagram of a scanner according to another embodiment of the present invention;

Fig. 25 is a schematic diagram showing the process of combining three partial images scanned by the scanner according to another embodiment of the present invention: (a) scan three partial images along directions A, B and C respectively, (b) preview of partial images aligning, (c) combining the partial images, and (d) a processed image of the partial images;

Figure 26 is a schematic side view of a scanner according to another embodiment of the present invention;

Fig. 27 is a schematic diagram showing the process of combining two partial images scanned by a scanner according to an embodiment of the present invention: (a) two partial images scanned along the first scanning direction A and the second scanning direction B respectively, (b) a combination of the partial images, and (c) a processed image of the partial images;

FIG. 28 is a schematic diagram of a scanner according to another embodiment of the present invention;

Fig. 29 is a schematic diagram showing the process of combining four partial images scanned by a scanner according to an embodiment of the present invention: (a) scan four partial images along directions A, B, C and D respectively, (b) partial images The pre-alignment of , (c) combination of part images, and (d) processed images of part images;

Figure 30 shows schematic side views of different embodiments of the scanner of the present invention: (a) one embodiment of the scanner, (b) another embodiment of the scanner, and (c) still another embodiment of the scanner Way;

Fig. 31 shows the glare elimination process of the scanner shown in Fig. 30(c).

FIG. 32 is a schematic diagram of a scanner according to an embodiment of the present invention;

Figure 33 is a schematic side view of a scanner according to yet another embodiment of the present invention;

Figure 34 is a schematic diagram of different embodiments of the scanner of the present invention: (a) a side view of one embodiment of the scanner, (b) a side view of another embodiment of the scanner, and (c) a further embodiment of the scanner Bottom view of the embodiment;

Fig. 35 is a schematic diagram of a scanner according to another embodiment of the present invention.

Detailed ways

Since numerous modifications and variations in the present invention will be apparent to those skilled in the art, the present invention is more particularly described in the following examples which are provided for illustration only. Various embodiments of the invention will now be described in detail. Referring to the drawings, like reference numerals indicate like parts throughout. As used in the specification and claims below, the meanings of "a", "an" and "the" include plural reference unless stated otherwise. Also, as used below in the specification and claims, unless otherwise stated, the meaning of "in" includes "in" and the meaning of "above (above)".

Embodiments of the present invention are described with reference to accompanying drawings 3-35. In accordance with the objects of the present invention, by way of specific and broad description, in one aspect, the present invention relates to a scanner for obtaining an image of an object located on an at least partially transparent platform. The objects include documents and the like.

3 and 4, a scanner 100 according to an embodiment of the present invention includes an at least partially transparent platform 101, a light source 301, a rotatable mirror 102, a fixed mirror 110, an image sensor 104, a condenser lens 317, and an image processing system 120.

In an exemplary embodiment, platform 101 has a first scan area 101a and a second scan area 101b. Each of the first scanning area 101a and the second scanning area 101b of the platform 101 has a first edge 101a1 , 101b1 and a second edge 101a2 , 101b2 respectively. The length and direction from the first edge 101a1 to the second edge 101a2 of the first scan area 101a define a first scan length _L1 and a first scan direction A, respectively. Furthermore, the length and direction from the first edge 101b1 to the second edge 101b2 of the second scanning area 101b define a second scanning length _L2 and a second scanning direction B, respectively. In one embodiment, the first scanning area 101a and the second scanning area 101b overlap to form an overlapping scanning area 101c with an overlapping scanning length _L3 . Therefore, the scanning length of the actual scanning area of the platform 101 is L=(L ₁ +L ₂ −L ₃ ). The scan length L represents the maximum image length of the target that the scanner 100 can capture. As shown in FIG. 3 , the platform 101 has a white area 201 formed at least partially around the edge portion 101 d of the platform 101 , and the white area 201 has a plurality of marks 203 . Each mark 203 is located at a predetermined position of the white area 201 . In one embodiment, the plurality of markings 203 are identifiable to the white area 201 and preferably have a simple geometric shape such as a cross. The platform 101 has a platform made of an at least partially transparent material. In one embodiment, the platform 101 has a glass plate or a transparent plastic plate.

A light source 301 is used to emit light 105 . As shown in Figure 4, the light source 301 is arranged such that when the light source emits a light beam 105, said light beam 105 is focused by a lens 319, turned by a mirror 306, and then the light beam is turned by a beam splitter 315 along a first direction 105a, thereby reaching a Rotating mirror 102 . The light source 301 may be one of lasers, fluorescent tubes, light emitting diode (hereinafter referred to as "LED") components, tungsten lamps, tungsten halogen lamps, halogen lamps, xenon lamps and any combination thereof. For fluorescent tubes and xenon lamps, a parabolic cover for collecting and collimating emitted light to a predetermined direction needs to be provided at a predetermined position. The xenon lamp has a visible light spectrum in the wavelength range of about 400 nm to about 700 nm and a high output power. For example, Hamamatsu's xenon short arc lamps, L2173 and L2193 (Hamamatsu Photonics, k.k., Hamamatsu, Japan) have an output power consumption of 35W. In one embodiment, the invention is practiced using a xenon lamp.

The rotatable mirror 102 is adapted to receive light 105 from a first direction 105a and reflect the light toward a second direction 105b for scanning a partial image of an object such as a document located on the platform 101, and the rotatable mirror 102 receives a partial image of the scanned object from a third direction 105c and reflects said partial image of the scanned object in a fourth direction 105d. In one embodiment, the third direction 105c is opposite to the second direction 105b, and the fourth direction 105d is opposite to the first direction 105a. In the embodiment shown in FIG. 4 , the partial image of the object reflected to the fourth direction 105 d is guided to the condenser lens 103 and then to the image sensor 104 via the beam splitter 315 . The rotatable mirror 102 includes a flat mirror with at least one reflective surface, and the reflective surface is coupled with an image processing system 120 . In one embodiment, the rotatable mirror 102 includes a polygon mirror. In implementation, as shown in FIG. 4 , the rotatable mirror 102 is rotated in a predetermined direction 106 at a constant angular velocity by a rotating device 107 such as a rotating motor. The angular velocity is adjustable.

On the optical path between the rotatable reflector 102 and the first scanning region 101a of the platform 101, a fixed reflector 110 is arranged, and the fixed reflector 110 is used to receive light reflected from the rotatable reflector 102 along the second direction 105b. and reflect the light received from the rotatable mirror 102 toward the first scanning area 101a of the platform 101 along the fifth direction 105e, thereby being used to scan the partial image of the target at the position 112; and the fixed mirror 110 is also used for A scanned partial image of the object at position 112 is received from a sixth direction 105f opposite the fifth direction 105e and is reflected towards the rotatable mirror 102 along a third direction 105c. In one embodiment, the fixed mirror 110 includes a plane mirror. In another embodiment, the fixed mirror 110 comprises a curved mirror.

The platform 101, the rotatable mirror 102 and the fixed mirror 110 are arranged so that when the rotatable mirror 102 is rotated, the rotatable mirror 102 makes the light in the second direction 105b along the first scanning direction A from the first scan area 101a One edge 101a1 to the second edge 101a2 and along the first scanning direction B from the first edge 101b1 to the second edge 101b2 of the second scanning area 101b sequentially scan the continuous partial image of the object, preferably only once in the rotatable mirror 102 The above-mentioned scanning can be realized in a full rotation. In this configuration, a first angle α ₁ is defined between the platform 101 and the optical path 109 connecting the lower edge 110 b of the fixed mirror 110 with the first edge 101 a 1 of the first scanning area 101 a. A second angle _α2 is defined between the at least partially transparent platform 101 and an optical path 109b connecting the intersection of the first and second directions 105a, 105b and the first edge 101b1 of the second scanning area 101b. In this embodiment, _{both α1} and _α2 are greater than the predetermined threshold angle α. In one embodiment, said predetermined threshold angle a corresponds to the worst case of image distortion of the scanner and in fact determines the minimum height of said scanner. Compared with the traditional scanner shown in Figs. 1 and 2, assuming that the scanner of the present invention and the conventional scanner all have the same scanning length and the same worst image distortion, that is, α ₁ =α ₂ =α ₀ =α, this The scanners shown in Figures 3 and 4 of the invention are more compact than conventional scanners such that H<H ₀ .

The image sensor 104 is used for receiving the scanned partial image of the object reflected by the rotatable mirror 102 and passing through the condenser lens 103 along the fourth direction 105d and outputting an electronic signal corresponding to the received scanned partial image of the object to the image processing system 120 . The image sensor 104 may be a line sensor, an area sensor or a combination thereof. For a scanner employing a linear image sensor, the line processing rate of the line sensor is essential to increase the scanning speed of the scanner. For example, μPD3747 and μPD8670 produced by NEC Corporation in Tokyo, Japan have 7400 pixels and a data output rate of 44MHz, converted into a row processing rate of 5.95KHz. These line sensors are used to achieve a scanning speed of one page per second at a resolution of 300dpi. The line processing rate of the sensor DALSAIT-P1-2048 from DALSA Company is 46KHz, and this sensor can be used to realize the scanning speed of two pages per second when the resolution is 600dpi. Other commercially available line sensors may also be used to practice the invention.

The image processing system 120 is configured to receive electronic signals from the image sensor 104 and record the electronic signals in an electronic format. The image processing system 120 combines the partial images recorded therein to form a substantially complete target image, which is equivalent to a full scan along the first scanning direction A and the second scanning direction B respectively. The image processing may be implemented by software or firmware, and may be performed by a computing device physically located inside or outside the scanner and in communication with the scanner. In one embodiment, the image processing system 120 includes a computer having a plurality of microprocessors and software packages installed thereon. Commercially available microprocessors such as the Silicon Optics sxW1/sxW1-LX (Silicon Optics inc., Salt Lake City, Utah) are fast enough for image processing in fractions of a second. A software package such as Halcon (MVTEC Software GmbH, Müchen, Germany) can be used to process the electronic signal received from the image sensor 104 into a complete target image. Other microprocessors, software packages, and custom software may also be used to implement the invention.

Referring to Figures 5 and 6, and referring specifically to Figure 5 first, the image processing system according to the first embodiment of the present invention performs the following steps: Step 130, preprocessing the image of the scanned part of the target received from the image sensor, so that the independent black pixels Whiten and blacken individual white pixels to extract a high-gradient image called the reference image. At step 132, the image is corrected using a reference image. The image correction step includes an image combination step 131 , an image shading intensity correction step 133 and an image distortion removal step 135 . The order of the image processing flow in the three steps 131 , 133 and 135 may be different in different implementations. In any order, partial images scanned from documents need to be combined, shading strength needs to be corrected, and image distortion needs to be removed. FIG. 5 shows one of several possible image processing flows within the scanner image processing system 120 shown in FIGS. 3 and 4 .

To reduce the time required for image processing, multiple markers located in the white area of platform 101 can be used to quickly, reliably and accurately combine partial images into a complete image, and the markers can also be used to time the angular coordinates of the rotating mirror and Eliminate image distortion. The standard whiteness of the white area can be used as a reference for shading intensity for shading correction. For example, as shown in FIG. 6, overlapping portions of scanned partial images 206 and 207 of common marks 204 and 205 near scanned partial images 206 and 207 are processed as reference images. Other parts can also be processed as reference images. There is a one-to-one correspondence between pixels in the extracted reference image and pixels in the original scanned partial image. A pixel in the reference image corresponds to a pixel in the original scanned image at the same location. Accurately aligning and combining this reference image and combination allows the original scan portion images to be aligned and combined. Image processing to align and combine the fiducial images to align and combine the original scan portion images can be performed quickly and reliably due to the compact and expected shape of the markers within the overlapping portions in the fiducial images. The aforementioned reference image is optional in the image processing system of the present invention. It is also possible to directly align and combine the scanned partial images of the target with multiple marks 203 on the white area 201 .

Shading correction is performed at the shading intensity correction step 133 . Since there are several partial images and different parts of the partial images are obtained at different viewing angles and at different distances. In one embodiment, shading intensity correction is performed based on shading data obtained by reading a reference white area having a plurality of marks as shown in FIG. 6 during signal reading from an original partial image. For reasons similar to those of shading variations, different parts of the scanned part image have different distortions and need to be removed in the image distortion removal step 135 . Finally, in the image trimming step 134, the image of the white area is clipped from the processed image to obtain a complete and desired target image.

With flatbed scanners, documents need to be manually placed on and removed from the scanner during image scanning, ie, there are interruptions in scanning. The scanning time of the scanner also includes the time to manually put the document on the scanner and remove the document from the scanner. Therefore, scanning a large number of documents will result in slower average scan speeds than scanning only a few pages. For example, a scanner according to an embodiment of the present invention can continuously scan a 20-page document within 10 seconds, or scan a 120-page document within 1 minute without scanning interruption. Therefore, the scanning speed of the scanner is 2 pages per second. However, if the scanner is manually operated for about 10 minutes, with an interruption of about 5 minutes, only 600 pages can be scanned in total. The overall, effective scan speed for high-volume document operations then becomes one page per second. Although the above numbers are hypothetical, they are sufficient to show that the scanning speed of a large number of scans is significantly lower than that of a document of a few pages. When temporarily storing scanned partial images in the memory buffer of an image processing system according to an embodiment of the present invention, the image processing system may use computer processing capabilities connected to the scanner via, for example, a USB port, or a wireless protocol, e.g. , the computer can also be used for purposes other than image processing and does not have to be dedicated to the scanner. Therefore, this further reduces the cost of the scanner.

To further reduce processing costs, the plurality of markers, as shown in Figure 6, preferably each have a simple and identical geometric shape, such as a cross. In this structure, as shown in FIG. 7 and FIG. 8, only some predetermined simple image signal patterns need to be stored in the image processing system, and these patterns are used for and from scanning lines such as scanning line 112 in FIG. 3 and FIG. The extracted actual signal is compared to identify the mark and identify the relative position of the scan line to the mark. The simple shape of the marker can reduce the workload of image recognition in the image processing system. Additionally, the markers can take different shapes.

During scanning, the rotational timing method of the rotatable mirror is very tolerant to geometric errors in the scanner structure due to manufacturing process inaccuracies or material deformation over years of use. Also, no motor (not shown) driving the rotatable mirror is required to coincide with the scanning step. The only requirement for the drive motor is to rotate at a steady angular velocity. Therefore, the image scanner of the present invention can be manufactured at low cost and is very durable.

Furthermore, for easy synchronization between physical mirror rotation and electronic image processing, the image can be read in more than one scan. This is possible because the rotation of the rotatable mirror enables faster scanning speeds than users expect. Therefore, it is still very fast for the user to scan the document using two or more scans. The first scan can be used to identify the position of the marker and to synchronize with the image processing. The second scan is used to actually read the image for processing. Since the rotatable mirror is rotated at a constant angular velocity, when the first scan is completed, the start time of the second scan can be easily and accurately calculated. The physical rotation of the rotatable mirror and the image processing of the scanned image can thus be synchronized in the second scan. Alternatively, if two scans are used to capture an image of a scanned document, both the first scan and the second scan can capture an image of the document. The image captured at the first scan can be used as a "pre-scan image", allowing the image processing unit to calculate the necessary adjustments for capturing the second scan. This approach has other benefits as well. The calculated light intensity adjustment can be used to change the luminous intensity of the second scan. For example, if an LED is used as a light source, since the LED can be turned on or off quickly, its light intensity can be changed rapidly by applying different supply voltages. Therefore, the image quality captured by the second scan can be improved.

In the present invention, relatively expensive stepping motors, complex mechanisms for angular coordinate and speed measurement and speed control are not required for the method for scan line positioning and image synthesis. The main requirement of the rotation motor of the rotatable mirror is that it can rotate at a constant speed. The task of synchronization between the angular coordinates of the rotatable mirror and the processing of the image scan is undertaken by an image processing system which localizes the position of the scan line by identifying the position of the marker. This structure uses a minimum number of physical components and has low demands on the precision of the manufacturing process of the inventive scanner.

Further details of employing simple marker shapes for efficient processing of scanned images are described below.

Referring now to FIGS. 7 and 8 , and first to FIG. 7 , there is shown a magnified image of the scanned area 101 with white area 201 , marker 209 and circled area 208 in FIG. 7( a ) with xy coordinates, respectively. The coordinates x _a , x _b , x _c , x _d and x _e are the positions of the scanning lines 210 respectively, and the width w ₃ is the width of the white area 201 . The width w ₁ is the width of one arm of the cross mark 209 . The width w ₂ is the lateral width of the cross mark 209 . Coordinates y ₁ , y ₂ , y ₃ , y ₄ , y ₅ and y ₆ are respectively the position of the outer edge of the white area 201, the position of the outer edge of the horizontal arm of the mark 209, and the position of the outer edge of the vertical arm of the mark 209 , the position of the inner edge of the vertical arm of the mark 209 , the position of the inner edge of the horizontal arm of the mark 209 , and the position of the inner edge of the white area 201 . 8(a)-8(e) respectively show the IY plane views of the scan lines _xa , _xb , _xc , _xd and _xe of FIG. 7, where I represents the shading intensity. Assuming image processing starts at position y ₁ and runs in the opposite direction of the y-axis, the processing time required to identify the scanline at position x _a is t(r _a ), where r _a is the starting position y ₁ and position y The number of pixels between ₄ and much less than the number of pixels on the entire scan line at position _xa . Since the image processing time is proportional to the number of pixels that need to be processed. Fewer pixels means shorter processing time. Specifically, as long as the system detects a signal extracted on a scan line from position y ₁ to distance y ₄ , the processing logic can recognize that the scan line does not hit the marker 209 shown in FIG. 7 .

As shown in FIG. 8 , since _rc < r _b < _ra , the processing time for identifying the scan line at position x _b or x _c is less than t(r _a ) on average. As shown in FIG. 7 , if the scan line 210 at the last position does not touch the mark 209 and is at the current position x _b of said scan line 210 , the process encounters a transition from white to black at a distance from y ₁ to y ₃ The signal intensity changes on the scan line 210 are shown as I _w to I _b in FIG. 8 , respectively. The processing logic then considers the scan line to be located at the upper end of the mark 209 as shown in FIG. 7 . The same decision logic can be used to determine the positions of the scan lines _xc , _xd , _xe .

The signal from the line sensor can also be configured to scan the local pattern itself, as the scan line traverses the entire scan area in the scan directions A and B shown in FIG. 3 . Due to the amount of image processing involved in the scanner of the present invention, the amount of computation is greatly reduced. The manner in which the position of the scan line 210 is detected using the mark 209 on the white area is shown in FIGS. 7 and 8 . When the scan line 210 is located at the position _xa , the entire scan line 210 is located in the white area. As shown in FIG. 8( a ), the image signal extracted on the scan line 210 located on the reflective surface of the white area 201 is white with an intensity _Iw .

The scan line 210 intersects the partial mark 209 when the scan line is at position _xb . As shown in FIG. 8( b ), the signal extracted on the scan line 210 located on the reflective surface of the white area 210 has a dark segment with a width equal to the width w ₁ of the mark 209 . At position x _c , the scan line 210 intersects the middle of the mark 209 . As shown in FIG. 8( c ), the signal extracted from the scan line 210 has a dark segment whose width w ₂ is equal to the entire mark width w ₂ shown in FIG. 7 . In FIGS. 7 and 8 , w ₃ represents the width of the white area 201 . As shown in FIG. 8( d ), at position x _d , when limited to the white region 201 , the signal extracted from scan line 210 has the same shape as at position x _b . In FIG. 8( e ), at position x _e , the signal extracted from scan line 210 has the same shape as at position x _a when confined to the white area 201 .

Since the position and shape of the markers are predetermined and stored in the image processing system, the markers can be used to quickly and reliably detect the x- and y-coordinates of any pixel on the scan line. Specifically, if the position of the scan line relative to the marker is known, the position of any pixel on the scan line can be determined by interpolation.

Once the coordinates of each pixel on the scanline and the coordinates of the scanline are determined, combining the multiple scanlines into a 2D image is a quick and straightforward process. Since the position and shape of the markers are known to the image processing system, the processing cost of forming the two-dimensional image is reduced.

Referring to FIG. 9 , two partial images 211 and 212 scanned by a scanner according to an embodiment of the present invention may be pre-aligned by an image processing system. The scanned portion image 211 has marker images 214 and 215 respectively corresponding to markers 204 and 205 located in respective white areas of the overlapping scan region 200 shown in FIG. 6 . The scanned portion image 212 has marker images 216 and 217 respectively corresponding to markers 204 and 205 located in respective white areas of the overlapping scan region 200 shown in FIG. 6 . Marker images 214 and 215 located in scanned portion image 211 and corresponding marker images 216 and 217 located in scanned portion image 212 are used to further align the two images 211 and 212 . In ideal alignment, marker images 214 and 216 should completely overlap and marker images 215 and 217 should also completely overlap. The simple geometry and strong color contrast of the marker with its surrounding white area allows for quick and reliable alignment of the partial images in the combining step 131 of FIG. 5 .

The partial image combination step can also be performed in the same way if the two partial images do not share a common marker such as markers 204 and 205 in FIG. 6 .

FIG. 10 also shows the combining steps for obtaining a pair of partial images from a single scanning cycle of a document in the scanner embodiment shown in FIGS. 3 and 4 . The image signals enter the image processing system in the sequence indicated by directions A and B as the scan proceeds. In other words, image scanning is performed in directions A and B, and scanning A and scanning B can be done during a single rotation of the rotatable mirror. According to a separate embodiment, during a single rotation of the rotatable mirror 102, A may be scanned before B or B may be scanned before A. As shown in FIG. 4 , scan A obtains an image through a rotatable mirror 102 and a plane mirror 110 . Thus, scan A produces a flipped image as shown in Figure 10(a). As shown in Figure 10(b), the flipped image needs to be flipped again, and then as shown in Figure 10(c), with the help of marks such as 1010 and 1020 at the junction of the white area and other parts obtained by scanning B Image composition. Finally, the image distortion is removed as shown in Fig. 10(d). Referring to FIG. 5 , distortion is eliminated in a distortion removal step 135 .

Referring first to FIG. 11 , referring to FIGS. 11-15 , there is shown a flowchart for one of three systems provided for the red-green-blue (hereinafter "RGB") mode in the shading intensity correction step 133 of FIG. 5 . According to the exemplary embodiment shown in FIG. 5 , the image signal output from the line sensor includes a two-dimensional partial image, which is combined into a complete scanned image of the document by the image combining step 131 of FIG. 5 . As shown in FIG. 11, the shading intensity correction step 133 includes image recognition processing 230, scan address generation processing 232, shading correction data lookup table (hereinafter referred to as "LUT") 234, shading correction data with coordinates 238, and shading correction data 238. When step 236. In one embodiment, image recognition process 230 is used to identify markers and their locations in the image to be processed. The locations of the markers are then used in a scan address generation process 232 to calculate the coordinates of each pixel in the image. The calculated coordinates are converted to addresses in the shading correction data LUT 234. The found shading correction data is sent together with the coordinates of the shading correction data 238 to the shading processing unit 236, where the required correction amount is calculated using the real shading data from the white area image and the shading intensity of the image is corrected accordingly.

The shading correction is performed on the scanned image signal of the document in the shading correction process 133 shown in FIG. 11 . The calculations for the shading correction steps are further shown in FIGS. 12 to 15 . FIG. 12( a ) is a perspective view of a scanner according to the embodiment of the present invention shown in FIGS. 3 and 4 . Meanwhile, FIG. 12( b ) schematically shows the shading intensity in the scanning area 101 of the scanner. Figures 12(a) and 12(b) are aligned to show the relationship between areas 206 and 207 of scan area 101 and corresponding shading intensity surfaces 256 and 257. The shading strength surface 256 represents the shading strength in the area 206 including the white area at the edge of 206 . The shading strength surface 257 represents the shading strength of the region 207 including the white region at the edge of 207 .

The computational steps involved in the shading process 236 step shown in FIG. 11 are further illustrated in FIG. 13 . As an example, curve 140 represents shading correction data stored in graphical form in shading correction data LUT 234 of FIG. 11 . The numerical values of these data have significance only relative to each other. In other words, the curve is a curve shape dedicated to shading correction data useful in shading correction processing. When the shading correction data of the curve is looked up in 234 of FIG. 11 , the end value of the curve matches the actual shading intensity 142 of the image of the white area in FIG. 13 , which graphically represents the brightness data of the image of the white area. The actual shading intensity is measured by the image obtained by the white area. All values on the curve 140 are adjusted according to the endpoints of the curve while keeping the shape of the curve. In this way, the curve is converted from 140 to 144 using the image of the white area, and 144 represents the adjusted shading intensity correction data.

As shown in FIG. 13 , all individual lookup curves are adjusted based on the shading intensity values at both ends of the curves and curves such as 1310 represent the adjusted shading intensity correction data, which form the surface 146 of FIG. 13 . In the shading correction data LUT 234 in FIG. 11 , the shading correction curve can be searched along the x direction or the y direction. Curved surfaces need to be further processed to "smooth". As shown in FIG. 13, the shading intensity on the edge 148 of the curved surface 146 is the actual shading data obtained from the image of the white area. The final step in this shading data correction process is to calculate the difference between the expected shading intensity surface 150 and the measured shading intensity surface 146 for each location (x, y) and use these differences to correct the shading intensity of the original document image.

Fig. 14 is a cross-sectional view of the measured shading intensity surface and the corrected shading intensity surface projected on the I-X plane. Specifically, Fig. 14(a) shows the luminance before shading correction. Curves 265 and 267 represent the shading intensity of the partial images 263 and 264 . Figure 14(b) shows the shading intensity after shading correction.

FIG. 15 is an I-Y plan view, in which solid line 2010 represents the projection of part of the measured shading intensity surface 256 in FIG. 12 , and dashed line 2020 in FIG. 15 represents the corrected shading intensity surface.

FIG. 16 is a schematic diagram of a distortion removal process using white areas with marks according to an embodiment of the present invention. Also shown are data for generating, storing, and using grids in an image processing system, methods for mapping grids to distorted original images, and methods for image processing to remove distortions to obtain undistorted scanned documents The original image. In order to simplify the description, in the following description, the term "pixel" is used to indicate a small area on the image about the size of a pixel and may or may not correspond to an actual pixel on the image. Moreover, for convenience, in the following discussion, a processed image with no or little distortion is referred to as a processed image, and an image subject to distortion removal processing is referred to as an original image.

Fig. 16(a) shows how a coordinate grid is generated. Usually, the lines constituting the grid in the original image are not straight lines, such as 1610. There are various ways to structure and store the data of these lines in the image processing system. The following description provides an example of how these lines are constructed and stored.

In order to obtain the data to construct the coordinate grid, reverse engineering methods can be used. A reference whiteboard 1620 different from the white area on the scanner, such as that shown in FIG. An image of the plate is obtained by using the scanner of the present invention. The coordinate grid of the original image is shown in FIG. 16( a ). The data used to construct the grid may be obtained by comparing the coordinates of the grid located on a whiteboard such as 1630 that has not been scanned with the image coordinates of a whiteboard such as 1640 that has been scanned.

The coordinate grid data is stored in an image processing system by one of several methods. These methods are well known to those skilled in the art. For example, one approach is to store the coordinates of a sufficient number of points on the line of the coordinate grid. Another approach is to store only the coordinates of certain points lying on the coordinate grid with the data used to construct the curve (such as a spline) to fit these points. All of these methods can be used to practice the present invention.

Figures 16(c) and 16(d) show the interpolation steps for forming a coordinate grid located on the image of the scanned area by means of marks located in the white area of the scanner. Using a coordinate grid helps the image processing system remove distortion quickly. In the first step, as shown in Fig. 16(c), multiple lines are used to connect the markers located on the distorted image to form a rough coordinate grid such as 1605, which has a relatively large grid size and anchors on each mark. Specifically, the coordinates of pixels located on a coarse coordinate grid such as 1605 are calculated by interpolation based on known marker coordinates. Then, in the second step, as shown in FIG. 16(d), a fine coordinate grid such as 1615 is constructed based on the coordinates of pixels located on the coarse coordinate grid such as 1605 constructed in the first step. Specifically, the coordinates of pixels located on the fine coordinate grid are calculated by interpolating from the known coordinates of the pixels on the coarse coordinate grid already constructed in the first step. In order to reduce the size of the memory required for the calculations, only the coordinates of some selected points on the grid are actually calculated and stored. Repeat this process to obtain a finer grid based on the grid constructed in the previous step. The following discussion assumes that two levels of the coordinate grid may provide sufficient precision for performing distortion cancellation in various embodiments of the invention.

During the distortion removal step, the coordinates of the markers in the processed image (target of the transformation) are first calculated. Then, calculate the pixel coordinates in the processed image where some selected points of the fine coordinate grid are located on the coarse coordinate grid. Thereafter, computing the processed image from the original image is a fast process. Fig. 16(e) and Fig. 16(f) show the calculation steps for the distortion removal of any pixel P within the coordinate grid, which pixel is not located at the edge of the coordinate grid. Although not required, it is conveniently assumed that the grid in Figure 16(e) is small enough for a given level of distortion such that the four edges of the second level (thinner, or smaller size) Basically assumed to be a straight line. Determine the original image by the coordinates of the four corner points 00, 01, 10 and 11 and the ratios d ₁ /d ₁ ', d ₂ /d ₂ ', d ₃ /d ₃ ' and d ₄ /d ₄ ' The location of pixel P. As shown in Figure 16(f), the coordinates and ratios c ₁ /c ₁ ', c ₂ /c ₂ ', c ₃ /c ₃ ' and c ₄ /c ₄ ' determine the position of the same pixel P in the processed image. This ratio is preserved before and after distortion removal, that is, d ₁ /d ₁ '=c ₁ /c ₁ ', d ₂ / ₂ '=c ₂ /c ₂ ', d ₃ /d ₃ '=c ₃ /c ₃ ', d ₄ /d ₄ '=c ₄ /c ₄ '. Therefore, after the coordinates of the four corner points in the processed image are calculated, the calculation amount for obtaining the coordinates of the pixel P in the processed image is very small. Since the data amount of the pixel coordinates in the grid is much larger than the data amount of the four corner point coordinates of the grid, the calculation of the distortion removal transformation is very efficient. Fig. 16(a) and Fig. 16(b) show the overall view of the partial image distortion removal step using the above processing method.

Figure 16(g) shows the distortion due to different tilt angles at different angular coordinates of the rotating mirror. This distortion can be removed using the grid method mentioned above together with other distortions such as shown in Fig. 16(a).

16(h) is an illustration of the step of mapping the coordinates of the stored coordinate grid to the distorted original image and the white area image and marks that have been extracted from the scanned document. The coordinate grid is stored using the method mentioned above, ie a sufficient number of points lying on said coordinate grid line (not the points for interpolation and the coefficients of the function) are stored. This step description only refers to a local area of the distorted image, but it should be understood that the same steps can be performed on the entire image. During scanning, an original image 400 is obtained by scanning an original document. The original image 400 includes images of markers 410, 415 and 417 located in white areas. Lines 420 , 422 , 424 , and 426 are the original image of the grid located on the original image 400 . Point 450 is a point located on a stored coordinate grid whose coordinates have been stored in the image processing system. It is assumed that the point 450 lying on the line 420 deviates from the line 420 of the original image 400 due to distortion occurring during the scanning process. Therefore, 450 needs to be "mapped" to 420. The first step of the corresponding process is to correspond the stored edge marker points 430, 435 and 437 to the actual positions of the markers 410, 415 and 417 on the original image 400 in the directions of the arrows shown in FIG. 16(h). When the correspondence is completed, all points located on the stored coordinate grid assumed to be at the position of the marker are mapped to the corresponding marker of the original image. In the second step, the stored coordinates of other points such as 440 and 445 assumed to be at the edge of the coordinate grid are corresponding to the appropriate positions on the original image at the edge of the coordinate grid 422 and 426 respectively. Since points 440 and 445 maintain their relative positions to points 430, 435 and 437, the corresponding step is based on the marked positions on the original image. In addition, as shown in FIG. 16( h), the point 450 located on the internal line of the stored coordinate grid is corresponding to the appropriate position on the line 420 of the original image coordinate grid (since the point 450 will remain consistent with the point The relative positions of 430, 435, 440, 445 and 437 remain unchanged) while maintaining the shape of the grid.

The multiple steps shown in Figures 16(a)-16(h) propose a method for storing coordinate grid data in an image processing system, for mapping a coordinate grid to a distorted original image, and for processing to remove distortion There are several ways to obtain undistorted original images of scanned documents. These methods can be performed quickly. In addition to the additional calculations required to construct the coordinate grid and map the coordinate grid to the distorted original image, the extensive calculations used to remove the distortion also include copying pixel data one by one from the location in the distorted original image to the undistorted The corresponding position of the processed image. For black and white images, the pixel data includes pixel coordinates and pixel brightness. For color images, the pixel data includes pixel coordinates and the brightness of the red, blue and green light of the pixel. As a rough estimate, if the image consists of 1028×768 pixels, distortion removal can be achieved by taking 786,000 pixel copy operations.

If the image processing work is carried out by software executed by a separate personal computer located outside but connected to the scanner, today's mid-range personal computers have a clock speed of about 1-3 GHz and can do so in fractions of a second for high-resolution images Complete the distortion removal process. If the scanner is equipped with a computing device with its own dedicated processor, since the execution of copying pixel data from the original image to the processed image is the type of operation that is easily performed in parallel, when the When image processing, the execution time of distortion removal can be further reduced.

The extra computation required to construct the coordinate grid and map it to the original image depends on the requirements for distortion removal. If it is assumed that the finest grid constructed based on the coarse grid is approximately a straight line, since the pixels are assumed to be assembled orderly and linearly in the coordinate grid, the larger the size of the finest grid, the straight line premise The greater the error introduced. Conversely, the larger the size of the finest grid, the less calculation required for constructing the coordinate grid and corresponding the coordinate grid to the original image. Therefore, the calculation of the distortion correction is faster.

An advantage of the above method using a grid of coordinates for implementing distortion removal is also that the errors are evenly distributed over the entire image so as not to significantly affect the quality of the processed image.

Algorithms and data for partial image combination, distortion removal and shading correction calculations can be fixed at the time of manufacture or turn up of the scanner. In addition, algorithms and data for correcting partial image combinations, distortion removal and shading correction calculations can be dynamically selected during each scan run based on the marker image and scanned or pre-scanned white areas. The second option uses more system resources and may result in a more expensive scanner. However, the second solution has the advantage that a scanner using the second solution can tolerate larger deformations, geometric errors, brightness variations than a scanner using the first solution. In both schemes, the calculation step uses the "standard white" of the white area as a reference and uses the position marked on it as the basis for the calculation.

Referring now to FIG. 17, a scanner according to an embodiment of the present invention is shown. The light source 301 emits light beams. The emitted light beam is focused by lens 319 and then reflected by mirrors 306 and 305 respectively. A mirror 305 is positioned close to the image path 105d. After reflection by mirror 305, the light beams having dimensions 105a1 and 105a2 bounding them travel in a path close to image path 105d, but in the opposite direction to 105d. After the light beam is reflected by 305, the light beam reaches the rotatable mirror 102 and the rotatable mirror 102 reflects the light beam and causes the light beam to propagate inside the boundaries 105b1 and 105b2 corresponding to the front boundaries 105a1 and 105a2 respectively, the light beam further reaches the reflection The mirror 110 and said mirror 110 reflect the light beam and cause it to propagate inside the boundaries 106b1 and 106b2 corresponding to the front boundaries 105b1 and 105b2, respectively. As shown in FIG. 17 , setting the width of the light beam, the proximity of the reflector 305 to the image optical path 105d and the angle of the reflector 305 makes the distance defined between 106b1 and 106b2 during the entire scanning of the original document The beam irradiates the area 307 on the scanning area 101 , the beam moves together with the scanning image line 112 and always covers the scanning image line 112 . In other words, an area at or near a scan line, such as 112, is illuminated during a scan during which the scan line is moving. The image of said 112 travels along 106a, and after being reflected by mirror 110, travels along 105c, and then is reflected by rotatable mirror 102 so that it travels along 105d and reaches optical sensor 104. The angle defined between 105b1 and 105c is equal to 180°+β. The angle defined between 105a1 and 105d is equal to 180°-β. The angle β has a value in the range of -15° to 15°. Preferably, the value of angle β is substantially close to zero.

FIG. 18 is a schematic diagram of scanner illumination according to an embodiment of the present invention. The illumination path 105b has a constant angle β with respect to the image path 105c during scanning. The two optical paths intersect each other only at a certain distance from the rotatable mirror 102 . This distance may be chosen as the distance between the rotatable mirror 102 and the point at which the illumination beam path encounters the scanning area at one end 1410 of the scanning area 101 . In other words, both the illumination light path and the image light path hit the scanning area at the same location 1410 while scanning occurs at one end of said scanning area 101 . While the scan is progressing towards the center 1450 of the scan area, the two optical paths hit the scan area at slightly different locations such as 1420 and 1430 . The closer the distance to the center of the scanning area, the farther the distance between the two optical paths reaching the scanning area 101 .

This phenomenon has two effects. The first effect is that scan lines located at or near 1410 can be better illuminated than locations such as 1420, 1430 and 1450 that are more centrally located in the scan area. The reason is that the illumination beam path hits the scan area at exactly the same location as the image beam path at one end of the scan area, such as at position 1410 . This effect helps to balance the illumination intensity for the scan area since the ends of the scan area such as locations 1410 require stronger illumination light than areas near the center of the scan area such as locations 1420 , 1430 and 1450 . The degree of this effect can be adjusted by changing the focus of the scanning beam and by changing the geometry of the various light sources, mirrors and lenses along the optical path. A more focused beam from the source produces a narrower illuminated area in the vicinity of the scan line on the scan area compared to a weakly focused beam. Thus, as described above, a more focused scanning light has a stronger illumination balancing effect, i.e., the intensity of illumination on a scan line when the rotatable mirror scans one end of the scan area such as position 1410 and when the rotatable mirror 102 scans The greater the difference between the illumination intensities of the scan lines when scanning the center of the region such as positions 1420, 1430 and 1450. A weaker focused scanning light has a weaker illumination balancing effect.

The second effect is that when the scan is at or near the center of the scan area, the optical path from the light source to the scan area and back to the optical sensor at any point during the scan according to the principles of geometric optics cannot reach the optical sensor. This means that when the portion of the image path between the rotatable mirror and the scanning area is approximately vertical, strong reflections from the document support material used to support the document, such as glass or any shiny surface on the document, will not Influx of said optical sensor. So there is little chance of spoiling the image quality. The degree of this effect can be adjusted by changing the focus of the scanning beam, changing the angular range of the scan, and changing the position of mirrors, light sources, sensors, etc.

On the other hand, in this configuration, the illumination path is never parallel to the image path during the entire scanning process. Assuming that "specular reflection" is defined as the reflected light from the document support glass of the scanner (or any other transparent material supporting the scanned document) having the same angle of incidence as the illuminating light but in the opposite direction, i.e. The normal line of the light reflection point is symmetrical, which can be determined by the relationship between the angle of incidence and the angle of reflection in geometric optics. When the specular reflection does not hit the rotatable mirror (which occurs when the scan is in a position away from the center area), it is unlikely that the specular reflection with strong, unwanted glare at the optical sensor will obscure the real image from the document. Image light condition. When the specular reflection hits the rotatable mirror (this occurs at some point when the image path is at or nearly perpendicular to the scan area), since the specular path is not parallel to the image path (see Figure 18), And because the optical path from the rotatable mirror to the optical sensor is long, the specular reflection cannot reach the optical sensor. In this way, possible glare due to specular reflections can be eliminated.

An alternative or supplementary solution to the above mentioned strong glare at approximately 90 degree angles is to set up two scanning illumination structures, if the scanning illumination is arranged in such a way that the beam never touches the scanning area close to or at the position of the scanning line surface and the beam is perpendicular to the scanning field (ie, at an angle of approximately 90 degrees to the scanning field), this completely eliminates glare from reflected light. However, this solution has certain disadvantages, as compared with the vertical size of the scanner, its scanning area is relatively small. The reason is that the part of the image path between the scanning area and the rotatable mirror is not perpendicular to the scanning area. In addition, anti-reflection coatings on all lenses and on the scanning area of the scanner can also reduce reflections and improve image quality.

The two illumination devices shown in Figures 4 and 17 provide focused light at and near scan lines such as 112 as the scan progresses. A cylindrical lens or lens system 319 positioned between the light source 301 and the first mirror 306 is used to focus the light emitted by the light source 301 and form a narrow beam of high intensity light around a scan line such as 112 . The high-intensity beam enables the scanner to perform high-speed scanning.

Compared with the illumination balance effect shown in FIG. 18 , alternative and/or supplementary schemes for adjusting the illumination intensity of the scanning area are described below. In order to compensate the effect of the optical path length variation and the angular variation of the optical path in the scanning area during scanning, the power source of the light source 301 can be arbitrarily modulated by the rotation angle of the rotatable mirror. More specifically, the control logic is executed in the following steps: (1) determine the angular coordinates of the rotatable mirror by identifying the marker position on the partial scan image in real time; (2) use the detected angular coordinates to find the required light intensity at that moment The correction data is then used to control the power of the power supply 301 . Generally, when the light path is long and the incident angle of the light path to the scanning area is large, the light source 301 needs more power. Otherwise, the light source 301 requires less power. The light intensity of the LED light source can be changed quickly by changing the power supply and this light source is suitable for this situation.

Yet another solution for replacing or compensating the above two methods for illumination balance is to set a light shield 1510 between the rotatable mirror 102 and the scanning area as shown in FIG. 19 . The gobo 1510 has variable transparency (not shown) at different angles during rotation of the rotatable mirror 102 . Different transparencies can be used at different angles to adjust the light intensity of the scanned area and the amount of reflected light 1910 that travels back to the optical sensor 104 .

Figure 20 presents a scanner according to an embodiment of the present invention. The image path (scan line starting at the scan area and ending at the optical sensor) and the illumination path (scan line starting at the light source and ending at the scan area) within the scanner are shown in the graph of FIG. 20 . Lenses 501 and 503 focus light from light source 507 into beams of narrow vertical distance and wide horizontal distance parallel to the scan lines. Reflector 505 concentrates light from light source 507 onto lens 503 . A focusing lens 521 focuses image light reflected from the scan lines and light directly from the rotatable mirror 102 onto the light sensor 522 . The stray light directly emitted by the opaque cover 531 from outside the scanning area 101 makes the light illuminating the scanning line on the scanning area 101 only come from the rotatable mirror 102 . In contrast to FIG. 4, the embodiment of FIG. 20 uses only one rotatable mirror, without additional mirrors such as plane mirror 110 of FIG. The use of a rotatable mirror produces a relatively small scanning area compared to the vertical distance of the scanner. Or, in other words, for a scan area of the same size, the vertical distance (thickness) of FIG. 20 is relatively large. The novelty of this embodiment of the invention lies in the illumination arrangement illustrated in Figure 17 and the use of white areas and marks. If necessary, the mask shown in FIG. 19 can be arranged on the upper part of the rotatable mirror so as to balance the light intensity of the entire scanning area.

As shown in FIG. 20, a fluorescent tube 507 is used as a light source. Fluorescent lamps are low cost, have a good white light spectrum, and are highly efficient. However, fluorescent lights can cause flicker due to the modulated current passing through the fluorescent tubes. Current fluorescent lamp technology employs high frequency electronic ballasts that can generate a modulated current at a frequency of approximately 25KHz to 40KHz. If the scan area length is 12 inches, the scan resolution is 600 dpi (dots per inch), and the scanner scans one page per second, then, in one second, the total number of scan lines is 7,200 and the total number of flickers is 40,000 (assuming fluorescent lights The blinking frequency is 40KHz). Each line is allocated 40000/7200 = 5.5 blinks. The flickering of fluorescent lights degrades the quality of scanned images. However, the random nature of the flicker distributed over the image scan line makes the flicker insignificant across the entire image. Other types of light sources may also be used, such as tungsten lamps, tungsten halogen lamps, xenon lamps, light emitting diodes, and the like. Typically these light sources do not have the flickering problem that fluorescent lamps have. Therefore, when using other types of light sources, the scanning speed can be higher than when using fluorescent lamps. Each light source has its own advantages and disadvantages. LEDs and the like can be switched on and off instantaneously, producing concentrated and directional light without generating a lot of heat. However, in general, LEDs are not as efficient as fluorescent lights and are not very bright. Xenon lamps have a good spectrum and are very bright, but require a high voltage power supply, which makes them expensive and bulky. Tungsten lamps are not very efficient and generate a lot of heat. Since tungsten lamps are point light sources, while fluorescent lamps and LEDs are line light sources, complex designs of reflectors and lenses are required to convert the point light into a linear beam. Tungsten-halogen lamps are only slightly more efficient than tungsten lamps and have similar problems as tungsten lamps.

Figure 21 presents a scanner according to an embodiment of the present invention. In this embodiment, two rotatable mirrors, two optical sensors, and two light sources are used to achieve a large scanning area while keeping the vertical distance of the scanner small. Together with the reflector 605, the light source 607, and the lenses 603, 601 produce an intense, narrow and focused beam of light parallel to the scan lines used for scanning. The beam is deflected by a rotatable mirror such as 102 and projected onto a scan area 101 around the scan line. The focusing lens 621 focuses the image light reflected from the scanning line located in the scanning area to the optical sensor 622 . The two partial images scanned by the two rotatable mirrors are combined by image processing software with the aid of markings on the white area, which is located around the scanning area. Precise mechanical synchronization is not required between the rotations of the two rotatable mirrors, as long as the two rotatable mirrors can scan the image in a reasonably short period of time, e.g. rotate, and the image processing system can capture the two partial images in time and process them. Two image paths are shown in FIG. 21 , starting from the scan area 101 and ending at the optical sensor 622 , and two illumination paths, starting from the light source 607 and ending at the scan area 101 . If necessary, a light shield as shown in FIG. 19 may be provided on the upper parts of the two rotatable mirrors between the rotatable mirrors and the scanning area, so as to balance the light intensity of the entire scanning area.

Figure 22 presents a scanner according to another embodiment of the present invention. In this embodiment, one rotatable mirror and two optical sensors are used. In order to achieve a large scanning area while keeping the vertical distance of the scanner small, two auxiliary mirrors 701 and 703 are used. Although not required, it is preferable to make the surfaces of the auxiliary mirrors 701 and 703 in a curved shape. Advantages of the curved shape of the auxiliary mirror include small mirror surface area, easy geometry for positioning the optical sensor and light source, and a short overall vertical distance for the scanner. If necessary, a light shield as shown in FIG. 19 can be provided between the rotatable reflector 102 and the two auxiliary reflectors 701 and 703 under the rotatable reflector to balance the light intensity of the entire scanning area. The use of an opaque cover 707 prevents stray light from the rotatable mirror 102 from being emitted directly outside the scanning area. This is done to protect the user's eyes from injury while the user is operating the scanner.

In FIG. 22 , it is easier for the image processing system to perform distortion removal, partial image combination and shading correction on the scanned image by determining the curvature of the curved auxiliary mirror graphically or manually. The shape of a curved mirror is not unique.

The curvature of the curved auxiliary mirror can also be determined mathematically by calculation. An example of this calculation is as follows. Referring to FIGS. 22 and 23 , point A is the center of rotation 2210 of the rotatable mirror 102 . The illuminating light reflects from point A and hits curved mirror surface F at point D. This assumption introduces a small error since the diameter of the rotatable mirror is not zero. This small error can be ignored for this discussion if the diameter of the rotatable mirror is small enough, ie, the rotatable mirror is "slender" enough. Assume that the projection of the surface F on the XY plane is represented by the function y=f(x). DB is the normal of curved mirror F at point D, which passes through point D and is perpendicular to the tangent f(x) at point D. Assuming angle BDC (2420) = angle ADB (2410) and the X-axis represents the scan area 101, the slope of AD is tan(θ ₁ )=f(x)/x and the slope of DB is tan(θ ₂ )=-1/ f'(x). The following relationship can be established among θ ₁ , θ ₂ , x and f(x):

1/tan(θ ₁ )−1/tan(θ ₂ )=x/f(x)+f′(x).

Moreover, from geometry,

1/tan(θ ₁ )−1/tan(θ ₂ )=AB/f(x).

Combining the above two equations, the following equation can be obtained,

AB/f(x)=x/f(x)+f'(x).

Rearranging the above equation,

AB=x+f'(x)f(x)

Then take derivatives on both sides of the obtained equation to obtain the following equation,

d(AB)/d(θ ₁ )=d(x)/d(θ ₁ )+f'(x)*df(x)/d(θ ₁ )+f(x)*df'(x)/ d(θ ₁ ).

The above differential equation can also be written as

d(AB)/d(θ ₁ )=d(x)/d(θ ₁ )+f(x)/x*df(x)/d(θ ₁ )+f(x)*d( f(x)/x)/d(θ ₁ ).

As a design purpose, as the rotatable mirror 102 is rotated, the angle _θ1 is changed and the line AD is rotated about the point A. The velocity v of motion of point C on the X axis is also a constant, i.e., dv/dθ ₁ =K ₁ , where K ₁ is a constant, or at least changes smoothly, i.e., d ² v/d ² θ ₁ =K ₂ , where K ₂ is a constant. The above-mentioned first objective is approximately that the moving speed of point B is a constant relative to θ ₁ , dAB/dθ ₁ =K ₃ , K ₃ is a constant, and AB represents the length of the connecting line from A to B.

In Figure 23, it is assumed that the angular velocity θ ₁ is constant, and if the moving velocity of point B is set to be constant then d ² (AB)/d ² θ ₁ = 0, then the above differential equation can be transformed into a second-order differential equation And solve it numerically to obtain the function f(x), which represents the desired shape of the curved mirror. Due to integration, the function f(x) will contain some unknown constants. In actual mechanical design, the constant can be determined by adjusting the positions of the endpoints of the curved mirrors 701 and 703 .

Figure 24 presents a scanner according to another embodiment of the present invention. Compared with FIG. 4 , the scanner in FIG. 24 has an additional flat mirror 160 in front of the flat mirror 110 . The size, shape and position of the plane mirror 160 are designed so that the mirror 160 does not block any part of the image path from the regions L ₁ and L ₃ to the rotatable mirror 102 . The image of the scanned document in area _L2 is firstly reflected by plane mirror 2510 and then by plane mirror 160 towards the rotatable mirror, and finally the image propagates to line sensor 104 . Additional mirrors 2510 and 160 extend the scan area length L ₂ -L ₅ . If L ₁ +L ₂ +L ₃ -L ₄ -L ₅ in Figure 24 is equal to L ₁ +L ₂ -L ₃ in Figure 4 (so that both scanners have the same scan length in their scan areas), And the minimum values of _α1 and _α2 in Fig. 4 are equal to the minimum values of _γ1 , _γ2 and _γ3 in Fig. 24 (so that only parts of the scanned document without further image processing are distorted by the same oblique degree limit), then the height H in the embodiment of FIG. 24 can be made smaller than the height H in the embodiment of FIG. 4 . Therefore, the additional flat mirrors 2510 and 160 reduce the height of the scanner, enabling the physical structure of the image scanning device of the present invention to be fabricated as a lower structure. Angle _β1 is the maximum viewing angle of the rotatable mirror when reflecting an image through planar lens 110. Angle _β2 is the maximum viewing angle of the rotatable mirror 102 when the image is reflected through planar lenses 2510 and 160, and angle _β3 is the maximum viewing angle of the rotatable mirror when the image is directly onto the rotatable mirror.

25(a)-25(d) illustrate the steps of combining and converting partial images obtained from the original document by the scanner of FIG. 24 into a complete front view of the original document. The scanning sequence of the three partial images is A, B, and C, and the scanning is performed along the directions of the arrows related to A, B, and C. Fig. 25(a) shows three partial images obtained by scanning. Flipping the scanned partial images of 176 and 177 and swapping positions for 175, 176 and 177, the result is shown in Figure 25(b). In FIG. 25(c), three partial images 175, 176, and 177 are combined. Combining the partial images is assisted by the images of position markers 180 , 181 , 182 and 183 . In particular, 180 denotes two images of the marker, one on partial image 175 and one on partial image 177 . The two marker images are obtained from the same marker located on the white area of the scanner. When combining partial images, two of said marker images are superimposed into one image. The same description applies to 181, 182 and 183 as well.

Figure 26 is a schematic diagram of an embodiment of the present invention. In order to extend the length of the scanning area 101 while maintaining a low height of the scanner, this embodiment employs two line sensors 104 and 2705 . α ₁ and α ₂ are the inclination angles seen from line sensors 104 and 2705 by reflecting rotatable mirror 102, respectively. The image of the scanned document located on the scanning area 101 is reflected by the rotatable mirror 102 and focused by the lens 103 , and then reaches the line sensors 104 and 2705 . The scanning process is realized by rotation of the rotatable mirror 102 . The viewing angle β ₁ corresponds to the maximum distance L ₁ on the scanning area that can be scanned using the line sensor 2705 . The viewing angle β ₂ corresponds to the maximum distance L ₂ over the scanning area that can be scanned using the line sensor 104 . Letters A and B in FIG. 26 denote scan directions corresponding to the directions of rotation of the rotatable mirror in FIG. 26 . When the height H of this embodiment is kept the same as that in FIG. 2 , a scan length of L ₁ +L ₂ -L ₃ larger than L can be manufactured. FIG. 27 shows the steps of processing the partial image obtained from the original document by the scanner of FIG. 26 into a complete front view of the original document.

Fig. 28 is a schematic diagram of yet another embodiment of the present invention. This embodiment also employs two line sensors 104 and 2905 . The difference between the embodiment in FIG. 26 and the embodiment in FIG. 28 is that two additional planar mirrors 110 and 2910 are used to further extend the length of the scanning area. The image parts of the scanned document located in the areas _L1 and _L4 on the scanning area 101 are reflected by the flat mirrors 110 and 2910 and then reach the rotatable mirror 102 . The image portion of the scanned document in areas _L2 and _L3 reaches the rotatable mirror 102 directly. Then, the image reflected from the rotatable mirror 102 reaches the line sensors 104 and 2905 through the lens 103 . Specifically, images from areas L ₁ and L ₂ reach line sensor 2905 and images from areas L ₃ and L ₄ reach line sensor 104 . Angle _β1 is the maximum viewing angle corresponding to distance _L1 over the scan area. Angle _β2 is the maximum viewing angle corresponding to distance _L2 . Angle _β3 is the maximum viewing angle corresponding to distance _L3 . Angle _β4 is the maximum viewing angle corresponding to distance _L4 . Distances L ₅ , L ₆ and L ₇ represent overlapping regions of adjacent regions L ₁ , L ₂ , L ₃ and L ₄ , respectively. γ ₁ , γ ₂ , γ ₃ and γ ₄ are the minimum inclination angles within the observed image in the regions L ₁ , L ₂ , L ₃ and L ₄ , respectively. If the minimum values of γ ₁ , γ ₂ , γ ₃ and γ ₄ in Fig. 28 are equal to the minimum values of α ₁ and α ₂ in Fig. 26 (ensure that when the document is scanned for the first time, scanners limit the same degree of distortion), and the length L ₁ +L ₂ +L ₃ +L ₄ -L ₅ -L ₆ -L ₇ of Figure 28 is equal to L ₁ +L ₂ -L ₃ of Figure 27 (to ensure Two scanners have the same scanning area length), then the height H in the embodiment of FIG. 28 can be lower than the height H in the embodiment of FIG. 26 . A, B, C, and D represent the propagation directions of the scanning lines located on the scanning area 101 . When the rotatable mirror is rotated, the entire scanning area can be scanned in the order of A, B, C, D along the directions indicated by the arrows associated with these letters. The scanning order in A, B, C, D is not mandatory and it is not necessary to complete a complete scan of the document in one rotation. For example, during one revolution, only range _L3 is scanned. During another rotation, _L1 is scanned. During one more revolution, _L2 is scanned. During another rotation, _L4 is scanned. In this embodiment, it takes four revolutions to achieve a full scan.

Figure 29 shows an example of the steps of combining partial images and removing distortion in an embodiment of the invention. Specifically, in Fig. 29(a), four distorted partial images are obtained by scanning. A, B, C, D indicate the scanning direction, that is, the direction in which the scanning signal enters the image processing system. In FIG. 29(b), two partial images corresponding to the scanning directions A and D are reversed. In Fig. 29(c), four partial images are combined using markers in the connected area. In Fig. 29(d), distortion is eliminated. As described above, the image processing order can be changed. For example, distortion may be removed in each partial image before combining the partial images.

Since the scanners of the present invention have a high scanning speed, they are suitable for scanning a large number of documents into image files in electronic format. When scanning many documents at high speed, the natural mode of operation is to scan documents without placing a covering over the scanning area. For greater operator comfort, it is desirable to reduce the amount of light emitted from outside the scanning area of the scanner. This is accomplished by turning on the scan light only when a scan is in progress and having some type of shading device to block the scan light between scans. Constructing the shading device or the switch that can operate the scanning light synchronously with the scanning process is well known to those skilled in the art, and details will not be described here.

In the first and second irradiation devices shown in FIGS. 4 and 17, when the optical path is perpendicular to the surface of the scanning area 101, the beam intended to irradiate the text along the scanning line can be reflected from the surface of the scanning area 101, and the reflected light can travel along the same path as the image path, and both enter the sensor. This will generate glare, where glare is defined as the phenomenon that the scanning beam produces strong reflected light on the surface of the scanning area due to the 90-degree angle between the scanning beam and the surface of the scanning area. Glare degrades the image quality of the initial scan.

In addition, there are various methods for avoiding occurrence of this phenomenon other than the method described in FIG. 18 . FIG. 30( a ) shows a method for avoiding glare in the first embodiment of the scanner of the present invention shown in FIGS. 3 and 4 . The maximum viewing angle _β2 can be defined so that when the image path is in the range indicated by _β2 , the optical path will never be perpendicular to the surface of the scanning area 101. The vertical position is indicated by a vertical line R. FIG. 30( b ) shows a method for avoiding glare in the second embodiment of the scanner of the present invention shown in FIG. 24 . The maximum viewing angle _β2 and the position and angle of the mirrors 2510 and 160 are set such that the maximum image path 3105 scanned through the mirror 2510 intersects the vertical line R. As shown in Figure 24, the area within _L2 can be scanned without the image path being perpendicular to the surface of the scanning area. Also, the combined region _L5 located between _L2 and _L3 is located to the right of said vertical line R. Therefore, under this structure, the optical path is not perpendicular to the surface of the scanning area at any position.

FIG. 30( c ) shows a schematic method for avoiding glare in the third embodiment of the scanner of the present invention shown in FIG. 26 . Since the rotatable mirror has a certain thickness, the travel path 515 of the scanning light reflected by the rotatable mirror 102 and the subsequent path from 515a to 515b to 515c will be in a vertical position at R1. The travel path 525 of the scanning light reflected by the rotatable mirror 102 and the subsequent path from 525a to 525b to 525c will be in a vertical position at R2. The span of combined region _L3 is made sufficient to contain both location _R1 and location _R2 . After the light 515 is reflected by the mirror 516 and the rotatable mirror 102 , the light 515 appears glare at the angular coordinate R ₁ . However, after the light 525 is reflected by the mirror 526 and the rotatable mirror 102, the light 525 propagates along 525d, and no glare occurs on the surface of the scanning area corresponding to the position _R1 .

The detailed steps for obtaining a glare-free image from the scanner shown in FIGS. 26 and 30(c) are further shown in FIG. Partial images 3205 and 3215 are image regions L ₁ and L ₂ in Fig. 30(c), respectively. Portion 3230 of partial image 3205 around position R ₁ contains glare and may be replaced by a corresponding portion 3210 located on 3215 . Likewise, portion 3220 of partial image 3215 around position R ₂ contains glare and may be replaced by a corresponding portion 3225 located on 3205 . After replacement, both partial images 3205 and 3215 are free of glare.

As shown in FIGS. 4 and 17 , changes in the image path length at various positions of the rotatable mirror result in less than ideal focus of the image of the lens 103 located on the image sensor 104 . Extending the length of the image path by collapsing the image path alleviates this problem. When extending the image path length, the ratio between the longest image path and the shortest image path is reduced. This ratio is always greater than 1. Suppose, if the longest image path is 900mm and the shortest image path is 700mm, the ratio is 900/700=1.286. If both the longest and shortest image paths are extended by 300mm, the ratio is reduced to (900+300)/(700+300)=1200/1000=1.2. Figure 32 shows a preferred embodiment of an extended image path. Taking the scanner of FIG. 24 as an example, as shown in FIG. 32, a mirror 77 is provided at a position where the line sensor 104 is to be used. The line sensor 104 receives the image reflected by the mirror 77 . A lens 103 focusing the reflected image is provided between the mirror 77 and the line sensor 104 . Thus, the longest and shortest image paths also extend the distance between the mirror 77 and the line sensor 104 .

Fig. 33 shows a schematic side view of a scanner according to yet another embodiment of the present invention. The scanner has a frame 3500 , a lens 3505 , an area sensor 3510 , a light source 3515 and a reflector 3520 . Line 3525 is the boundary of the area sensor 3510 viewing angle. The scanner has a white area containing markings located on the scanning area 3550 . The white areas with markers are used to combine the partial images obtained by the area sensor 3510 into a complete image. Since more than one area sensor is used, the "height" of this scanner can be reduced. A disadvantage of this design is that the light from the light source can be uncomfortable for the operator. To overcome this shortcoming, another version of this scanner was designed and schematically shown in Figure 34(a)-(c). As shown in FIG. 34( a ), this scanner has a frame 3600 , a lens 3610 , an area sensor 3620 , a light source 3630 , and a reflector 3640 . Since the operator usually stands at one side of the scanner embedded with the light source 3630 , it can be seen from the figure that the operator's eyes 3650 do not directly see the flash light emitted by the light source 3630 . Moreover, FIG. 34( a ) shows the scanning area 3660 and the boundary of the viewing angle of the area sensor 3670 . Fig. 34(b) shows the scanner structure viewed from B1 in Fig. 34(a). Fig. 34(c) shows the scanner structure from another perspective of B2 in Fig. 34(a). In FIG. 34(c), a white area 3680 and a mark 3690 are shown.

Fig. 35 shows another preferred embodiment of the invention regarding the use of fixed coordinates and markers. Markings 3710 are marked on the inner surface 3704 of the scanning area 101 of the scanner (the surface facing the mirror and lens, opposite the outer surface 3702). When there is more than one image capturing device 3720 , such as a lens, an optical sensor, a mirror for observing the document facing the outer surface 3702 of the scanning area 101 , the obtained partial images of the document overlap at least in the area 3750 . By adjusting the viewing angle of 3720, the image of mark 3720 is located on inner surface 3704 of scan area 101, between the "combined edges" of image paths 3730 and 3740, and thus is not in the final complete document processing image.

While several and alternative embodiments of the present invention have been shown, it should be understood that those skilled in the art may implement the invention without departing from what has been discussed and set forth in the foregoing specification and the essential scope specified in the following claims. Certain variations are made to the invention. Moreover, the embodiments are only used to illustrate the principles of the present invention and are not intended to limit the disclosed scope of the present invention.

Claims (54)

1, a kind of scanner (100) that is used for obtaining to be positioned at the image of the target on the partially transparent platform (101) at least, wherein partially transparent platform (101) has first scanning area (101a) and second scanning area (101b) at least, and first scanning area (101a) of the described platform of partially transparent at least (101) and each of described second scanning area (101b) have the first edge (101a1 respectively, 101b1) with the second edge (101a2,101b2), scanner (100) comprising:

A. be suitable for luminous light source (301);

B. rotatable mirror (102) is used for receiving and reflects described light from first direction (105a) light and to second direction (105b), is positioned at the parts of images of the target on the local at least transparent platform (101) with scanning; Described rotatable mirror also is used for receiving the parts of images of the target after the scanning and to the parts of images of the target of the four directions opposite with first direction (105a) after the described scanning of (105d) reflection from the third direction (105c) opposite with second direction (105b);

C. stationary mirror (110), be positioned at described rotatable mirror (102) and at least on the light path between first scanning area (101a) of partially transparent platform (101), described stationary mirror is used for receiving the light reflect from rotating mirror (102) along second direction (105b) and to the parts of images of the described light that receives from rotatable mirror (102) of first scanning area (101a) reflection of partially transparent platform (101) at least with the scanning target, and receive the parts of images of the target after the scanning, then with the parts of images of the target of third direction (105c) after rotatable mirror reflects described scanning;

D. imageing sensor (104) is used for from the parts of images of the target of four directions after (105d) receives scanning and the output electronic signal corresponding to the parts of images of the target after the scanning that receives;

E. image processing system (120) is used for receiving electronic signal and writing down described electronic signal with digital format from described imageing sensor (104);

Wherein, described rotatable mirror (102) and stationary mirror (110) are provided so that when described rotatable mirror (102) rotates, the second direction light (105b) that described rotatable mirror (102) changes light makes in an only omnidistance rotation of described rotatable mirror (102), first edge (101a1) of the corresponding light that reflects from stationary mirror (110) along the 5th direction (105e) along first direction of scanning (A) from first scanning area (101a) scans the continuous part image of target successively to second edge (101b2) to second edge (101a2) and first edge (101b1) along first direction of scanning (B) from second scanning area (101b)

Wherein, the parts of images that described image processing system (120) will write down is here combined to form complete substantially target image, is equivalent to respectively the full scan along first direction of scanning (A) and second direction of scanning (B).

2, according to the described scanner of claim 1, it is characterized in that, also comprise being positioned at the collector lens (103) on the light path between described rotatable mirror (102) and the described imageing sensor (104).

3, according to the described scanner of claim 1, it is characterized in that, also comprise the whirligig that is used to rotate described rotatable mirror (102).

According to the described scanner of claim 3, it is characterized in that 4, described rotatable mirror (102) comprises the level crossing with at least one reflecting surface.

According to the described scanner of claim 3, it is characterized in that 5, described rotatable mirror (102) comprises polygon mirror.

According to the described scanner of claim 1, it is characterized in that 6, described stationary mirror (110) comprises level crossing.

According to the described scanner of claim 1, it is characterized in that 7, described stationary mirror (110) comprises curved mirror.

According to the described scanner of claim 1, it is characterized in that 8, described imageing sensor (104) one of comprises in line sensor, area sensor and the combination thereof at least.

9, according to the described scanner of claim 1, it is characterized in that, described light source (301) comprise in laser, fluorescent tube, light-emitting diode component, tungsten lamp, halogen tungsten lamp, halogen lamp, xenon lamp and the combination in any thereof one of at least.

According to the described scanner of claim 1, it is characterized in that 10, the described platform of partially transparent at least (101), described rotatable mirror (102) and described stationary mirror (110) are arranged so that first jiao of α ₁Be defined as the described platform of partially transparent at least (101) and be connected and fixed the lower limb (110b) of catoptron (110) and the light path at first edge (101a1) of described first scanning area (101a) between the angle, and described first jiao of α ₁Greater than predetermined threshold α.

According to the described scanner of claim 10, it is characterized in that 11, described platform of partially transparent at least (101) and described rotatable mirror (102) are arranged so that and with second jiao of α ₂Be defined as the described platform of partially transparent at least (101) and connect described first direction (105a) and the light path at first edge (101b1) of the point of crossing of described second direction (105b) and described second scanning area (101b) between the angle, and described second jiao of α ₂Greater than predetermined threshold angle α.

According to the described scanner of claim 1, it is characterized in that 12, the described platform of partially transparent at least (101) comprises the platform that is made of the material of partially transparent at least.

According to the described scanner of claim 12, it is characterized in that 13, the described platform of partially transparent at least (101) comprises glass plate.

According to the described scanner of claim 12, it is characterized in that 14, the described platform of partially transparent at least (101) comprises transparent plastic sheet.

15, a kind of scanner that is used to obtain to be positioned at the image of the target on the partially transparent platform at least, wherein, at least the partially transparent platform has first scanning area and second scanning area at least, each of described first scanning area and described second scanning area has first edge and second edge respectively, and described scanner comprises:

A. be suitable for luminous at least one light source;

B. at least one rotatable mirror, be used to receive from the light of first direction and to second direction and reflect described light is positioned at the target on the described platform of partially transparent at least with scanning parts of images, receive the parts of images of the target after the scanning and to the parts of images of the target of four directions after the described scanning of reflection from third direction;

C. at least one imageing sensor is used for from the parts of images of the target of described four directions after receiving scanning and the output electronic signal corresponding to the parts of images of the target after the scanning that receives; And

D. image processing system is used for receiving electronic signal and writing down described electronic signal with digital format from described imageing sensor;

Wherein said at least one light source, at least one rotatable mirror and at least one imageing sensor are arranged so that described first direction and four directions are to limiting first jiao of 180 °-β, and described second direction and described third direction limit second jiao of 180 °+β, wherein β is the value in-15 ° to 15 ° scopes, and when described at least one rotatable mirror rotates, thereby described at least one rotatable mirror make the change of second direction of described light make only first edge of light along first direction of scanning (A) from first scanning area described in the omnidistance rotation at described at least one rotatable mirror to second edge and first edge along second direction of scanning (B) from second scanning area scan the continuous part image of described target successively to second edge

The described image processing system combination parts of images of record here is equivalent to respectively the full scan along first direction of scanning (A) and second direction of scanning (B) to form complete basically target image.

16, according to the described scanner of claim 15, it is characterized in that, also comprise the collector lens on light path between described at least one rotatable mirror and the described at least one imageing sensor.

17, according to the described scanner of claim 15, it is characterized in that, also comprise the whirligig that is used to rotate at least one rotatable mirror.

According to the described scanner of claim 17, it is characterized in that 18, described at least one rotatable mirror comprises the level crossing with at least one reflecting surface.

According to the described scanner of claim 17, it is characterized in that 19, described at least one rotatable mirror comprises polygon mirror.

According to the described scanner of claim 15, it is characterized in that 20, described at least one imageing sensor one of comprises in line sensor, area sensor and the combination thereof at least.

According to the described scanner of claim 15, it is characterized in that 21, the described platform of partially transparent at least comprises the flat board that is made of partially transparent material at least.

According to the described scanner of claim 21, it is characterized in that 22, the described platform of partially transparent at least comprises glass plate.

According to the described scanner of claim 21, it is characterized in that 23, the described platform of partially transparent at least comprises transparent plastic sheet.

24, a kind of method that is used to obtain to be positioned at the image of the target on the partially transparent platform at least, wherein have at least the first scanning area and second scanning area, comprise step by a plurality of marginal portions described platform of partially transparent at least of qualification and the described platform of partially transparent at least:

A. the marginal portion to small part around to the small part transparent platform forms the white portion with a plurality of marks, and described each mark is arranged on the precalculated position of described white portion;

B. respectively from the continuous part image of first scanning area and the described target of the second scanning area sequential scanning, the image that one of comprises in a plurality of marks at least of each continuous part image wherein; And

Thereby c. adopt a plurality of marks image one of at least be arranged in each continuous part image to form complete substantially target image with combination continuous part image corresponding to the full scan of first scanning area and second scanning area as benchmark.

25, according to the described method of claim 24, it is characterized in that the standard basis white that suits of at least a portion in the described white portion.

According to the described method of claim 24, it is characterized in that 26, each mark in described a plurality of marks all can be discerned from described white portion.

According to the described method of claim 24, it is characterized in that 27, described employing step also comprises the step of proofreading and correct formed target image.

28, according to the described method of claim 27, it is characterized in that, obtain the step of described target image thereby described employing step also comprises the image of the image of the white portion that prunes away respectively the target image after proofreading and correct and a plurality of marks.

According to the described method of claim 24, it is characterized in that 29, the described platform of partially transparent at least comprises the flat board that is made of partially transparent material at least.

According to the described method of claim 29, it is characterized in that 30, the described platform of partially transparent at least comprises glass plate.

According to the described method of claim 29, it is characterized in that 31, the described platform of partially transparent at least comprises transparent plastic sheet.

32, a kind of scanner that is used to obtain to be positioned at the target image on the partially transparent platform at least, wherein have at least the first scanning area and second scanning area, comprising by a plurality of marginal portions described platform of partially transparent at least of qualification and the described platform of partially transparent at least:

A. the white portion that has a plurality of marks, the marginal portion to small part around to the small part transparent platform forms, and each mark is arranged on the precalculated position of described white portion;

B. optical devices are used for respectively the continuous part image from first scanning area and the second scanning area sequential scanning target, wherein the image that one of comprises in a plurality of marks at least of each continuous part image;

C. image processing system, thus the image that one of is used for using in a plurality of marks of each continuous part image at least forms complete substantially target image corresponding to the full scan of first scanning area and second scanning area as benchmark with combination continuous part image.

33, according to the described scanner of claim 32, it is characterized in that the standard basis white that suits of at least a portion in the described white portion.

According to the described scanner of claim 32, it is characterized in that 34, each mark in described a plurality of marks all can be discerned from described white portion.

According to the described scanner of claim 32, it is characterized in that 35, described image processing system comprises controller.

36, according to the described scanner of claim 35, it is characterized in that, thereby described controller is also carried out the step that the image of the image of the white portion that prunes away respectively the image of the described target that proofread and correct to form and the target image after proofreading and correct and a plurality of marks obtains described target image.

According to the described scanner of claim 32, it is characterized in that 37, described optical devices comprise at least one imageing sensor.

According to the described scanner of claim 37, it is characterized in that 38, described at least one imageing sensor one of comprises in line sensor, area sensor and the combination thereof at least.

According to the described scanner of claim 32, it is characterized in that 39, the described platform of partially transparent at least comprises the flat board that is made of partially transparent material at least.

According to the described scanner of claim 39, it is characterized in that 40, the described platform of partially transparent at least comprises glass plate.

According to the described scanner of claim 39, it is characterized in that 41, the described platform of partially transparent at least comprises transparent plastic sheet.

42, a kind of method that is used to obtain to be positioned at the image of the target on the partially transparent platform at least, wherein, the described platform of partially transparent at least has a plurality of scanning areas, comprises step:

A. sequential scanning is respectively from the continuous part image of the target in each scanning area of described a plurality of scanning areas; And

B. make up described continuous part image with the complete substantially image of formation corresponding to the target of the full scan of a plurality of scanning areas.

According to the described method of claim 42, it is characterized in that 43, the described platform of partially transparent at least also has a plurality of marks that are set in the precalculated position respectively.

44, according to the described method of claim 43, it is characterized in that the image that each continuous part image one of comprises in described a plurality of mark at least.

According to the described method of claim 44, it is characterized in that 45, the image that described combination step one of comprises in described a plurality of marks of employing in each continuous part image at least is as the step of benchmark.

46, a kind of scanner that is used to obtain to be positioned at the target image on the partially transparent platform at least, the wherein said platform of partially transparent at least has a plurality of scanning areas, comprises step:

A. optical devices are used for sequential scanning respectively from the continuous part image of the target in described each scanning area of a plurality of scanning areas; And

B. treatment of picture device forms basic complete object image corresponding to the full scan of a plurality of scanning areas thereby be used for making up the continuous part image that receives from optical devices.

According to the described scanner of claim 46, it is characterized in that 47, described optical devices comprise at least one imageing sensor.

According to the described scanner of claim 47, it is characterized in that 48, described at least one imageing sensor one of comprises in line sensor, area sensor and the combination thereof at least.

According to the described scanner of claim 46, it is characterized in that 49, the described platform of partially transparent at least also has a plurality of marks that are separately positioned on the precalculated position.

According to the described scanner of claim 49, it is characterized in that 50, each continuous part image comprises as one of at least image in described a plurality of marks of benchmark.

According to the described scanner of claim 46, it is characterized in that 51, described disposal system comprises controller.

According to the described scanner of claim 46, it is characterized in that 52, the described platform of partially transparent at least comprises partially transparent plastic plate at least.

According to the described scanner of claim 52, it is characterized in that 53, the described platform of partially transparent at least comprises glass plate.

According to the described scanner of claim 52, it is characterized in that 54, the described platform of partially transparent at least comprises transparent plastic sheet.

Images