JP2007295210A - Image processing apparatus, image processing method, image processing program, and recording medium recording the same - Google Patents

Abstract

An image processing apparatus capable of improving the execution speed of image inclination correction by improving the content of line segment extraction processing in an image.
In an image processing apparatus for correcting the inclination of an image, an image input means for inputting image data, a line segment extraction means for extracting a line segment group in the image, and the line segment and a predetermined reference direction. An angle detection means for detecting an angle; and an image rotation means for rotating the image based on a detection result of the angle detection means, wherein the line segment extraction means is an edge region group consisting of edge regions of the image The image processing apparatus includes an edge region extracting unit for extracting a run length and a run length extracting unit for extracting a run length group composed of run lengths having a predetermined length or more based on the edge region group.
[Selection] Figure 3

Description

  The present invention relates to an image processing apparatus, an image processing method, an image processing program, and a recording medium on which the image processing program is recorded.

  With recent advances in hardware, various information electronic distribution services targeting not only PCs but also PDAs, mobile phones, and the like are expanding.

  For example, when manga is distributed electronically, content is created as image data. In this case, on the basis of the existing proofread books, manual operations such as (1) document scanner input, (2) image inclination correction, (3) image shaping by various image processing, (4) frame cutout / effect addition, etc. It is general to follow the creation process up to editing and (5) content output. In this process, except for manual editing (4), the process is automated and the content is created at a low cost.

  In the above-described series of processes, the image inclination correction process (2) is one of important processes. In particular, in the case of commercial content, it is necessary to correct the inclination with high accuracy in order to maintain the quality.

  In response to such a request, the conventional technology for automatically correcting the tilt of the input image has already been realized in various methods as a pre-processing function of an electronic filing system or OCR [Optical Character Reader]. ing. Moreover, as what specialized in the manga, the thing currently disclosed by patent document 1, for example is mentioned.

The prior art disclosed in Patent Document 1 provides processing for extracting edge pixels of an image, connecting edge pixels, assigning a direction code to the connected edge pixel row, and integrating the direction codes. In this way, horizontal / vertical line segments in an image mainly corresponding to the manga frame lines are extracted, the inclination of the document is determined from the inclination angles of the extracted plurality of line segments, and the inclination is corrected.
JP 2002-92606 A

  However, according to this conventional technique, when edge pixels are connected, a process of connecting the first unconnected edge pixel, ie, “pixel”, when the neighboring pixels of the target edge pixel are viewed in a predetermined rotation direction (clockwise in the conventional technique). Unit processing "is repeated. Therefore, there has been a problem that the processing load is relatively large and the execution speed is reduced.

  In view of the above problems, an object of the present invention is to provide an image processing apparatus capable of improving the execution speed of image inclination correction by improving the content of line segment extraction processing in an image.

  In order to achieve the above object, an image processing apparatus according to the present invention comprises an image input means for inputting image data and a line segment extraction means for extracting a line segment group in the image in the image processing apparatus for correcting the inclination of the image. Angle detection means for detecting an angle between the line segment and a predetermined reference direction, and image rotation means for rotating the image based on a detection result of the angle detection means, and the line segment extraction means An edge region extraction means for extracting an edge region group consisting of edge regions of the image, and a run length extraction means for extracting a run length group consisting of run lengths of a predetermined length or more based on the edge region group; It is set as the structure (1st structure) which has.

  According to this configuration, it is possible to extract a line segment group used for image inclination correction based on a run length group extracted in advance. For this reason, the execution speed of image inclination correction can be improved as compared with the conventional technique in which edge pixels are connected in units of one pixel and line segments are extracted. In addition, since the edge region extraction means for extracting the edge region group is provided, the run length group extraction process can be easily performed.

  Here, the “line segment” is an aggregate of pixels in an image, and an angle with a reference direction (for example, a horizontal direction or a vertical direction) can be detected based on the aggregate state. A “line segment group” refers to a set of line segments (including a single segment group).

  “Edge” refers to a pixel existing at the boundary when pixels having different attributes (for example, black and white) are adjacent to each other in the scanning direction, and “edge region” includes the vicinity of the edge. A set of pixels in a certain area is referred to, and an “edge area group” refers to a set of edge areas (including a single area). The “run length” is an aggregate of pixels that are continuous in a certain scanning direction. The “run length group” refers to a set of run lengths (including a single case).

  In the first configuration, the line segment extraction unit includes an integration unit that integrates run lengths having a predetermined connection relationship in the run length group, and the line segment is obtained by performing the integration. It is good also as a structure (2nd structure) to extract.

  According to this configuration, since the line segments are extracted by integrating run lengths having a predetermined connection relationship, the extraction process can be executed uniformly and reliably. The content of the “connection relationship” is not limited as long as it meets the purpose of extracting line segments.

  In the first or second configuration, the edge region extraction unit includes an edge extraction unit that extracts an edge group including the edges of the image, and an expansion unit that expands the image (third configuration). Configuration).

  According to this configuration, the edge group is extracted from the image, and each edge can be expanded by the expansion means. Therefore, the edge region group can be appropriately extracted. Here, “expansion” means that a pixel having a certain attribute is recognized as the same attribute for pixels in the vicinity of the pixel (for example, four pixels in contact with each other and eight surrounding pixels).

  In the first or second configuration, the edge region extraction unit includes an edge extraction unit that extracts an edge group including the edges of the image, and extracts the pixels in the vicinity of the edge, thereby extracting the edge region. It is good also as a structure (4th structure) which extracts a group.

  According to this configuration, the edge group is extracted from the image, and the edge region group is extracted by extracting the pixels in the vicinity of the edge. Therefore, the processing can be appropriately performed. The “pixels in the vicinity of the edge” here may be, for example, 4 pixels in contact with the edge pixel or 8 pixels surrounding the edge pixel, and as long as the appropriate line segment group is extracted later, The content is not limited.

  Further, in any one of the second to fourth configurations, the integration unit may calculate the run length when one run length and another run length are in contact with each other in the main scanning direction or the sub scanning direction. It is good also as a structure (5th structure) to integrate.

  According to this configuration, as a result of the integration of run lengths, it is possible to extract a line segment without interruption. Therefore, the angle detection process using the line segment can be easily performed. The “main scanning direction” here is a scanning direction related to run length extraction. Hereinafter, a direction orthogonal to the “main scanning direction” on the image plane is referred to as a “sub-scanning direction”.

  In any one of the second to fifth configurations, the line segment extraction unit performs the integration using reference range information based on the coordinates of the run length in the sub-scanning direction (sixth configuration). It is good.

  According to this configuration, since the integration process is performed using the reference range information (information for determining a range for checking the presence or absence of the connection relationship for a certain run length), the efficiency of the process can be improved, and the inclination of the image Correction can be performed at a higher speed. That is, for the purpose of extracting line segments, the integration targets for a certain run length are often limited to run lengths close in the sub-scanning direction. Therefore, by using reference range information, for example, by eliminating run lengths that are apparently not connected, from the investigation object in advance, the efficiency of the integration process is improved.

  Further, in any one of the second to sixth configurations, the line segment extraction unit performs a process of classifying the run lengths adjacent to each other in the sub-scanning direction among the run lengths in the run length group. It is good also as a structure (7th structure) which performs about a run length and performs the said integration based on this classification result.

  According to this configuration, based on the classification result, only those adjacent to each run length can be investigated as a connection relationship, so that the integration process can be made extremely efficient. Here, the “sub-scanning direction” refers to a sub-scanning direction related to run-length extraction.

  An image processing apparatus according to the present invention is an image processing apparatus that corrects an inclination of an image, an image input unit that inputs image data, a line segment extraction unit that extracts a line segment in the image, and the line segment. An angle detection unit that detects an angle with a predetermined reference direction and an image rotation unit that rotates the image based on the detection result of the angle detection unit (eighth configuration).

  The image processing method according to the present invention is an image processing method for correcting the inclination of an image, wherein an image input step for inputting image data, a line segment extraction step for extracting a line segment group in the image, and the line An angle detection step for detecting an angle between the minute and a predetermined reference direction, and an image rotation step for rotating the image based on the detection result (a ninth method).

  In the ninth method, the line segment extracting step includes an edge extracting step for extracting an edge group consisting of edges of the image, and an edge for extracting an edge region group by extracting pixels in the vicinity of the edge. A region extraction step and a run length extraction step for extracting a run length group consisting of run lengths having a predetermined length or more based on the edge region group, and in the run length group, a run having a predetermined connection relationship is included. A method (tenth method) for extracting the line segment group by integrating the lengths may be used.

  According to this method, in the line segment extraction step, the line segment for detecting the angle with the predetermined reference direction is appropriately selected through the extraction of the edge group, the extraction of the edge region group, and the extraction of the run length group. Can be extracted.

  An image processing program for operating the image processing apparatus according to any one of the first to eighth configurations, wherein the image processing program causes a computer to function as each of the above-described means. Image processing can be executed by a computer equipped with appropriate hardware. A computer-readable recording medium on which the image processing program is recorded is also useful.

  According to the configuration of the present invention, it is possible to extract a line segment group used for image inclination correction based on a run length group extracted in advance. For this reason, the execution speed of image inclination correction can be improved as compared with the conventional technique in which edge pixels are connected in units of one pixel and line segments are extracted. In addition, since the edge region extraction means for extracting the edge region group is provided, the run length group extraction process can be easily performed.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a block diagram showing a hardware configuration of an image processing apparatus according to the present invention.
Referring to FIG. 1, an image processing apparatus according to the present invention includes a control unit 101 that controls the entire image processing apparatus, a memory unit 102 that stores data, an input processing unit 103, an input device 104, The line segment extraction unit 105, the angle detection unit 106, and the rotation processing unit 107 are connected to each other through a bus 108.

  The control unit 101 generally corresponds to a CPU [Central Processing Unit], but may be configured by some kind of hardware. In the present application, a block that controls the entire image processing apparatus is referred to as a control unit, and its configuration is not limited.

  The memory unit 102 is generally composed of a ROM [Read Only Memory] and a RAM [Random Access Memory], but may be only a RAM, for example. In the present application, a block that holds various data of the image processing apparatus is called a memory unit, and its configuration is not limited.

  The input processing unit 103 is connected to the input device 104, binarizes image data input from the input device, and stores it in the memory unit 102 via the bus 108. The input device 104 is generally an image input device such as a scanner or a camera. In the present application, a device that inputs image data is referred to as an input device, and a block that processes the input data is referred to as an input processing unit, and the configuration thereof is not limited.

  The line segment extraction unit 105 is a block that extracts a line segment from the input binary image, the angle detection unit 106 is a block that detects the angle of the line segment and the image, and the rotation processing unit 107 is designated. A block that rotates an image based on an angle. The line segment extraction unit 105, the angle detection unit 106, and the rotation processing unit 107 can be implemented by a program using a CPU or by some kind of hardware, and the configuration thereof is not limited.

  FIG. 4 is a configuration diagram of the memory unit (102 in FIG. 1) of the image processing apparatus according to the present embodiment. The memory unit 102 includes an input image buffer 401, a binary image buffer 402, a line segment image buffer 403, an edge image buffer 404, a line segment coordinate buffer 405, a run length buffer 406, a candidate line segment buffer 407, and a fast search table buffer 408. Composed.

  Among these, the image buffers 401 to 404 normally store image data composed of a header portion and a data portion. Usually, the header portion includes a portion that stores at least the number of horizontal pixels and the number of vertical pixels. In general, there are a large number of such image data formats, and there are various configurations of the header portion and the data portion, but the configurations are not limited in the present application.

  FIG. 5 is a configuration diagram for one line segment data in the line segment coordinate buffer (405 in FIG. 4) in the memory unit 102 of the image processing apparatus according to the present embodiment.

  A plurality of line segment data is extracted by a line segment extraction unit (105 in FIG. 1). For each piece of line segment data, a line segment direction flag 501, a start X coordinate 502, a start Y coordinate 503, an end X coordinate 504, an end Y coordinate 505, and a line segment angle 506 are configured. The line segment direction flag 501 stores a flag value indicating the direction of the line segment. In the present embodiment, “1” is stored for a horizontal line segment, and “0” is stored for a vertical line segment. . Reference numerals 502 to 505 store circumscribed rectangular coordinates of the line segment. In the present application, the horizontal direction is represented by X and the vertical direction is represented by Y. The line segment angle 506 contains an angle with a reference direction (for example, the horizontal direction) detected for the line segment.

  FIG. 6 is a configuration diagram for one run-length data in the run-length buffer (406 in FIG. 4) in the memory unit (102 in FIG. 1) of the image processing apparatus according to the present embodiment.

  A plurality of run length data are extracted by a line segment extraction unit (105 in FIG. 1). One run length data is composed of an integrated label 601, a run direction flag 602, a sub-scanning direction coordinate 603, a main scanning direction start coordinate 604, and a main scanning direction end coordinate 605.

  The same label value is given to the integrated label 601 by the integrated run length data. The run direction flag 602 stores a flag value indicating the run direction. In this embodiment, “1” is stored for a horizontal run and “0” is stored for a vertical run. The sub-scanning direction coordinate 603 is given a coordinate in the sub-scanning direction orthogonal to the run direction as the main scanning direction. In the case of a horizontal run, the vertical coordinate is stored, and in the case of a vertical run, the horizontal coordinate is stored. In 604 and 605, the start coordinates and end coordinates of the run with the run direction as the main scanning direction are stored, respectively.

  FIG. 7 is a configuration diagram for one candidate line segment data in the candidate line segment buffer (407 in FIG. 4) of the memory unit (102 in FIG. 1) of the image processing apparatus according to the present embodiment.

  A plurality of candidate line segment data are extracted by a line segment extraction unit (105 in FIG. 1). One candidate line segment data is composed of an integrated label 701, a line segment direction flag 702, a start X coordinate 703, a start Y coordinate 704, an end X coordinate 705, an end Y coordinate 705, an evaluation value 707, and an adoption flag 708. Is done.

  The integrated label 701 is assigned a unique value for each candidate line segment data. 702 to 706 are information having the same meaning as 501 to 505 in FIG. The evaluation value 707 stores the evaluation value of the candidate line segment data. The adoption flag 708 is a flag indicating whether or not the candidate line segment data has been adopted, and “1” is entered when it is adopted and “0” is entered when it is not adopted. In addition, the adopted candidate line segment data 702 to 706 are copied to 501 to 505.

  FIG. 8 is a configuration diagram for one high-speed search table data in the high-speed search table buffer (408 in FIG. 4) in the memory unit (102 in FIG. 1) of the image processing apparatus according to the present embodiment.

  The sub-scanning direction coordinate 801 stores the sub-scanning direction coordinate where the run length exists. The start index 802 includes the first run-length data index having the sub-scanning coordinates stored in 801 in the run-length buffer, and the end index 803 includes the sub-length stored in 801 in the run-length buffer. Each last run-length data index having a scan coordinate is stored.

  The block diagram and each data structure according to the present invention have been described above. Hereinafter, an embodiment of the present invention will be described based on a flowchart. FIG. 2 is an overall flowchart of the image processing apparatus according to the present invention. The overall processing of the present invention will be described with reference to this figure.

  Image data is input from the input device 104 by the input processing unit 103. The input image data is stored in the input image buffer 401 (S201). The input processing unit 103 reads the contents of the input image buffer 401 and performs binarization processing (S202). In the binarization process, a certain threshold is defined for monochrome data (data in which one pixel is a value of 0 to 255 and expressed in multi-tone, for example, “1” for black, white In this case, the process is “0”.

  For general input image data, variations in the data format such as color, monochrome, binary, etc. are conceivable, but in the case of color, it may be converted into monochrome once and binarized. There are various methods such as discriminant analysis for binarization, but the binary method is not limited in the present invention.

  The binary image data is stored in the binary image buffer 402 (S203). The line segment extraction unit 105 performs line segment extraction based on the binary image data stored in the binary image buffer 402. This result is stored in the line segment image buffer 403 and the line segment coordinate buffer 405 (S204). The contents of the line segment extraction process will be described in detail again.

  The angle detection unit 106 detects an angle for each line segment based on circumscribed rectangular coordinates (502 to 505) stored in the line segment image buffer 403 and the line segment coordinate buffer 405. The detected angle is stored as a line segment angle 506 (S205).

  Note that the angle detection method for line segments is, for example, rotating a line segment image at every minute angle to obtain the peripheral part (black pixel histogram) in the main scanning direction and minimizing the variance of the peripheral distribution. A method of detecting the angle as a line segment angle is conceivable, but the method is not limited in the present invention.

  The angle detection unit 106 determines the angle of the entire image based on the line segment angle 506 for each line segment (206). As this image angle determination method, for example, the weighted average method of Equation 1 described in Patent Document 1 can be considered, but the method is not limited in the present invention.

  The control unit 101 determines whether the angle detection process has converged based on the image angle determined by the angle detection unit 106 (S207). If converged, the process branches to S209, and if not converged, the process branches to S208. If the angle detection process has not converged, the binary image is rotated based on the determined angle, new binary image data is created, and the process returns to S203.

  If the angle detection process has converged, the input image data stored in the input image buffer 401 is rotated based on the determined angle, and the inclination correction process ends (209). Note that the image rotation method may be simple affine transformation, linear interpolation, cubic interpolation, or the like, but the method is not limited in the present invention. The above is the description of the overall processing of the present invention.

  FIG. 3 is a flowchart of the line segment extraction unit of the image processing apparatus according to the present invention, which details the line segment extraction 204 in the overall flowchart of FIG. The line segment extraction process, which is a feature of the present invention, will be described below with reference to FIG. FIG. 9 is an example of input image data after binarization targeted for line segment extraction processing of the image processing apparatus according to the present invention. As can be seen from FIG. 9, the input image is inclined counterclockwise.

  The line segment extraction unit 105 performs edge image extraction (S 301) and expansion processing (S 302) on the input image FIG. 9 to obtain an edge region group, and this information is stored in the edge image buffer 404. The edge image extraction method is, for example, a place where scanning in the horizontal and vertical directions changes from a white pixel (a pixel having a pixel value of 0 in the present invention) to a black pixel (a pixel having a pixel value of 1 in the present invention). However, in the present invention, the method is not limited.

  In addition, as a method of expansion processing, all pixels in the image are sequentially scanned, and when the pixel of interest is black (edge pixel), a method of converting all the four or eight neighboring pixels into black pixels can be considered. In the present invention, the method is not limited. FIG. 10 shows an image obtained as a result of performing edge image extraction and expansion processing on the input image shown in FIG. That is, the dilation process is a process of generating an edge region group by extracting pixels in the vicinity of the edge and giving the extracted pixel the same attribute as the edge.

  The line segment extraction unit 105 scans the edge image buffer 404 horizontally and vertically, extracts only a long run length, and stores the result in the run length buffer 406 (S303). The run length is an aggregate of pixels that are continuous in the main scanning direction.

  The data shown in FIG. 6 is created for each run length and stored in the run length buffer 406. In addition, the long run length may be, for example, a method using a threshold value set according to the image size. In the present application, the method is not particularly limited, but in the present invention, only a run length longer than that is extracted with a value obtained by dividing the vertical or horizontal size of the image by 30 as a threshold value.

  FIG. 11 is a diagram illustrating a result of extracting only a long horizontal / vertical run length from the image of FIG. At this stage, the data of the run length buffer 406 shown in FIG. 6 is set as shown in Table 1 below for each extracted run length data.

  Further, FIG. 14 shows more specifically the relationship between the run length and the run length buffer 406 at this stage. Reference numeral 1401 in FIG. 14 denotes an image. The origin is the upper left. There are three coordinates x0, xs, and xe in the horizontal direction, and ys, ye, and y0 in the vertical direction. Reference numeral 1402 denotes a vertical run length, and reference numeral 1403 denotes a horizontal run length.

  In the case of 1402 vertical run length, the sub-scanning direction coordinate is x0, the start coordinate in the main scanning (run) direction is ys, and the end coordinate is ye. In the case of the 1043 horizontal run length, the sub-scanning direction coordinate is y0, the main scanning (run) direction start coordinate is xs, and the end coordinate is xe. The run-length buffer 406 in this case has the contents shown in the lower table of FIG.

  Next, the line segment extraction unit 105 scans the horizontal run and the vertical run in order, and in adjacent runs (runs that exist above and below in the case of horizontal runs, runs that exist on the left and right in the case of vertical runs), The runs are integrated by a method of assigning the same integrated label to those having a connection relationship (S304). When examining the connection relationship, it is determined that the connection is not established if the following conditions are satisfied.

Condition 1:
(Starting coordinate in the main scanning direction of the target run) −1 ≧ (coordinate in the main scanning direction of the integration candidate run)
Condition 2:
(End coordinate in the main scanning direction of the target run) + 1 ≦ (Starting coordinate in the main scanning direction of the integration candidate run)

  Based on the above determination, when run lengths are in contact with each other in the main scanning direction or the sub-scanning direction (including a case where a corner is shared), it is assumed that they have a connection relationship. It is assumed that it does not have. Note that the conditions relating to the determination of the connection relationship may be other contents as long as they meet the purpose of extracting an appropriate line segment.

  The same integration label as that of the target run is assigned to the integration candidate run that does not satisfy both condition 1 and condition 2. This determination is repeated until no new integration occurs, and run integration is performed. As a result of the integration process, the same integration label value (a natural number starting from 1) is assigned to the integration label 601 of the run-length buffer 406 to runs determined to have a connection relationship.

  Next, the line segment extraction unit 105 scans the run length buffer 406 that has been integrated, and creates a candidate line segment buffer 407 for run lengths that have the same integrated label (S304). At this stage, the data of the candidate line segment buffer 407 shown in FIG. 7 is set as shown in Table 2 below for each extracted candidate line segment data.

  Next, the candidate line segment buffer 407 is scanned by the line segment extracting unit 105, and an evaluation value is calculated based on the following equation 1 (S305).

Formula 1: (Evaluation value) = (Length in the main scanning direction) ² ÷ (Length in the sub scanning direction)

  This expression 1 is an evaluation value that increases as the elongated line segment increases. The evaluation value for each line segment is stored in the evaluation value 707 of the corresponding candidate line segment buffer 407. Next, this evaluation value 707 is scanned, and only a certain number (10 in the present invention) is extracted as an adopted line segment in order from the largest. For those determined to be adopted, “1” indicating adoption is stored in the adoption flag 708 of the corresponding candidate line segment buffer 407.

  Next, the line segment extraction unit 105 scans the candidate line segment buffer 407, creates a line segment image for only the employed line segment whose adoption flag 708 is “1”, and stores it in the line segment image buffer 403. (S306). FIG. 12 is a diagram of images stored in the line segment image buffer 403 as a result of line segment integration and adoption line segment determination based on the run length extraction result of FIG.

  Next, the line segment extraction unit 105 scans the candidate line segment buffer 407 and copies only the employed line segment for which the adoption flag 708 is “1” to the line segment coordinate buffer 405 ( S307). In this case, the line segment direction flag 702 is copied to 501, the start X coordinate 703 is copied to 502, the start Y coordinate 704 is copied to 503, the end X coordinate 705 is copied to 504, and the end Y coordinate 706 is copied to 505.

  The above is the line segment extraction processing of the image processing apparatus according to the present invention. As described above, edge pixels are connected in units of one pixel in the prior art. However, in the present invention, unit processing that integrates a plurality of pixels called run lengths can be performed, so that high speed can be realized. In addition, the run length can be easily extracted, and in that case, further increase in speed can be expected.

  However, in the line segment extraction 304 in the flowchart of FIG. 3, the run length buffer 406 is searched for all run lengths that are connected to the target run length, and redundant processing is performed in this portion. included. Eventually, since the connection relationship is judged by Condition 1 and Condition 2 for adjacent runs (runs that exist above and below in the case of horizontal runs, or runs that exist on the left and right in the case of vertical runs) Only the run length having the sub-scanning direction coordinates of ± 1 with respect to the sub-scanning direction coordinates need be targeted, and it is redundant to perform full scanning. The present invention provides means for solving this problem and further increasing the speed.

  FIG. 15 is a detailed flowchart of the line segment extraction (horizontal / vertical run integration) 304 of the line segment extraction unit 105 according to the present invention. FIG. 16 is a diagram showing the relationship between the run length and the run length buffer 406. First, the principle of speeding up will be described with reference to FIG.

  FIG. 16 shows only the horizontal run length for the sake of convenience, but the principle is the same for the vertical run length. Therefore, in the following description, without distinguishing between horizontal and vertical, the run direction is referred to as the main scanning direction, and the direction orthogonal to the run direction is referred to as the sub-scanning direction.

  In FIG. 16, 1601 indicates an image, and 1602 indicates a run-length buffer. The run length is shifted by one coordinate in the sub-scanning direction, and a run is extracted in the main scanning direction with respect to one sub-scanning direction coordinate. For this reason, the run lengths are arranged in the order of the sub-scanning direction when viewed globally. However, there may be no run length or a plurality of run lengths for one sub-scanning direction coordinate. This relationship is shown in FIG. 16, and the run length in the image 1601 is indicated by a colored rectangle (the hatched rectangle for the target run length), and the corresponding storage location in the run length buffer 406 is indicated by a dashed arrow.

  All the run length buffer elements 1602 need not be evaluated with respect to the target run length, and have a sub scan direction coordinate of ± 1 with respect to the sub scan direction coordinate of the target run length (sub It can be seen that only the object adjacent to the scanning direction is required. Specifically, in FIG. 16, this range is only between 1603 and 1604.

  In other words, among the run lengths in the run length group, a process of classifying adjacent run lengths in the sub-scanning direction is performed for each run length in the run length group. Then, by using information related to the classification result (reference range information based on the run-length coordinates in the sub-scanning direction), it is determined whether or not there is a connection relationship between the run lengths. The speed-up method of line segment extraction (integration of horizontal and vertical run lengths) of the present invention pays attention to this content, and will be described below with reference to FIG.

  First, the high-speed search table 408 is created (S1501). In the high-speed search table 408, one piece of data is the data shown in FIG. 8, all elements of the run length buffer 406 are scanned, and two or more run lengths having the same run direction flag and the same sub-scanning direction coordinate are scanned. Then, the index having the smaller index number in the run length buffer 406 is stored in the start index 802 and the index having the larger index number in the run length buffer 406 is stored in the end index 803 in order.

  If no run length exists for a certain sub-scanning direction coordinate, the data of FIG. 8 is not created, and if only one run length exists, the start index 802 and the end index 803 The values of are the same. (This is an index number in the run-length buffer 406 indicating that one run-length.)

  Next, various variables are cleared to 0 (S1502). This variable can be a memory variable, a register, or a hardware counter. In the present invention, three variables “integrated label”, “integrated flag”, and “current integrated label” are used.

  Next, an attention run length is set (S1503). This increments the index number in the run-length buffer 406 one by one, searches for the one having the run direction flag 602 that is the current integration target, and sets it as the target run-length.

  If the target run length has already been given an integrated label (if the integrated label is not “−1” indicating unintegrated), the value is set in the “current integrated label” variable. In the case of “−1”, the value of the “integrated label” variable is set to the “current integrated label” variable, and the value of the “integrated label” variable is incremented. In addition, an integrated label value is set in the integrated label 601 of the target run length.

  Next, a scan start / end index is acquired (S1504). The start index scans the high-speed search table 408, and the sub-scanning direction coordinate 801 is a start index 802 whose coordinate value is “−1” of the sub-scanning direction coordinate 603 of the target run length. The search table 408 is scanned, and the sub-scanning direction coordinate 801 is obtained as the end index 802 of the coordinate value of “+1” of the sub-scanning direction coordinate 603 of the target run length.

  In FIG. 16, when the hatched rectangle is the target run length, the start index is acquired as 1603 and the end index is acquired as 1604. If the start index cannot be set, there is no run length to be integrated into “(target run length sub-scanning direction coordinate) −1”, and therefore the start index is set to “(target run length index number) +1”. If the end index cannot be set, the end index is set to “(target run length index number) −1” for the same reason.

  Next, each run-length data in the run-length buffer 406 indicated by the start index to the end index is scanned in order, and it is determined whether or not the target run length can be integrated under the condition 1 and the condition 2. The value of the “current integrated label” variable is set in the integrated label 601 of the run length, and the value of the “integrated flag” variable is incremented (S1505).

  It is determined whether or not the run length of interest ends, and if it still exists, the process branches to S1503. If there is no more, the process branches to S1507 (S1506). It is determined whether the “integration flag variable” is “0”. If it is not “0”, the process branches to S1502. In the case of “0”, it is determined that the integration process is completed (S1507).

  As described above, the image processing apparatus according to the present invention provides means for processing tilt correction at high speed. FIG. 13 shows an image obtained as a result of correcting the tilt of the input image shown in FIG. 9 by the image processing apparatus according to the present invention. In particular, in the case of a copying machine or the like that requires high-speed processing, the image processing apparatus according to the present invention is effectively used.

  Furthermore, an image processing program for operating the image processing apparatus having the above-described configuration can be provided. With such a program, it is possible to cause a computer equipped with appropriate hardware to perform image processing that can enjoy the above-described advantages.

  Such a program is stored in a computer-readable recording medium such as a flexible disk attached to the computer, a CD-ROM [Compact Disc_Read Only Memory], a ROM [Read Only Memory], a RAM [Random Access Memory], and a memory card. It is also possible to provide a program by recording them. A program can also be provided by downloading via a network.

  The provided program product is installed in a program storage unit such as a hard disk and executed. The program product includes the program itself and a recording medium on which the program is recorded. The embodiment disclosed this time should be considered as illustrative in all points and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.

  The image processing apparatus, image processing method, image processing program, and recording medium of the present invention can be applied to a copying machine that requires high-speed processing.

It is a block diagram which shows the hardware constitutions of the image processing apparatus concerning this invention. 1 is an overall flowchart of an image processing apparatus according to the present invention. It is a flowchart of the line segment extraction part of the image processing apparatus concerning this invention. It is a block diagram of the memory part (102 of FIG. 1) of the image processing apparatus concerning this invention. It is a block diagram with respect to one line segment data of the line segment coordinate buffer (405 of FIG. 4) of the memory part (102 of FIG. 1) of the image processing apparatus concerning this invention. It is a block diagram with respect to one run length data of the run length buffer (406 of FIG. 4) of the memory part (102 of FIG. 1) of the image processing apparatus concerning this invention. It is a block diagram with respect to one candidate line segment data of the candidate line segment buffer (407 of FIG. 4) of the memory part (102 of FIG. 1) of the image processing apparatus concerning this invention. FIG. 5 is a configuration diagram for one high-speed search table data in a high-speed search table buffer (408 in FIG. 4) of a memory unit (102 in FIG. 1) of the image processing apparatus according to the present invention. It is an example of input image data after binarization targeted for line segment extraction processing of the image processing apparatus according to the present invention. FIG. 10 is an image obtained as a result of edge image extraction and expansion processing performed on the input image of FIG. 9 by the image processing apparatus according to the present invention. FIG. 11 is a diagram illustrating a result of the image processing apparatus according to the present invention extracting only long horizontal and vertical run lengths from the image of FIG. 10. FIG. 12 is a diagram of an image stored in a line segment image buffer 403 as a result of line segment integration and adoption line segment determination based on the run length extraction result of FIG. 11. FIG. 10 is an image obtained by correcting the tilt of the input image of FIG. 9 by the image processing apparatus according to the present invention. FIG. 6 is a diagram showing the relationship between horizontal / vertical run lengths and run length buffers 406; 6 is a detailed flowchart of line segment extraction (horizontal / vertical run integration) 304 of the line segment extraction unit 105 according to the present invention. It is a related figure of run length and run length buffer 406.

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 Control part 102 Memory part 103 Input processing part 104 Input device 105 Line segment extraction part 106 Angle detection part 107 Rotation processing part 108 Bus 401 Input image buffer 402 Binary image buffer 403 Line segment image buffer 404 Edge image buffer 405 Line segment coordinate Buffer 406 Run length buffer 407 Candidate line segment buffer 408 Fast search table buffer 1401 Image 1402 Vertical run length 1403 Horizontal run length 1601 Image 1602 Run length buffer

Claims (12)

In an image processing apparatus that corrects the tilt of an image,
Image input means for inputting image data;
Line segment extraction means for extracting line segments in the image;
Angle detection means for detecting an angle between the line segment and a predetermined reference direction;
Image rotation means for rotating the image based on the detection result of the angle detection means,
The line segment extracting means includes
Edge region extraction means for extracting an edge region group consisting of edge regions of the image;
An image processing apparatus comprising: a run length extracting unit that extracts a run length group including run lengths having a predetermined length or more based on the edge region group. The line segment extracting means includes
In the run length group, having an integration means for integrating run lengths in a predetermined connection relationship,
The image processing apparatus according to claim 1, wherein the line segment is extracted by performing the integration. The edge region extraction means includes
Edge extraction means for extracting an edge group consisting of edges of the image;
The image processing apparatus according to claim 1, further comprising an expansion unit configured to expand an image. The edge region extraction means includes
Edge extraction means for extracting an edge group consisting of edges of the image,
The image processing apparatus according to claim 1, wherein the edge region group is extracted by extracting pixels in the vicinity of the edge. The integration means includes
5. The run length according to claim 2, wherein when one run length and another run length are in contact with each other in the main scanning direction or the sub-scanning direction, these run lengths are integrated. Image processing apparatus. The line segment extracting means includes
The image processing apparatus according to claim 2, wherein the integration is performed using reference range information based on coordinates in the sub-scanning direction of the run length. The line segment extracting means includes
A process for classifying adjacent run lengths in the run length group in the sub-scanning direction is performed for each run length in the run length group,
The image processing apparatus according to claim 2, wherein the integration is performed based on the classification result. In an image processing apparatus that corrects the tilt of an image,
Image input means for inputting image data;
Line segment extraction means for extracting line segments in the image;
Angle detection means for detecting an angle between the line segment and a predetermined reference direction;
Image rotation means for rotating the image based on the detection result of the angle detection means;
An image processing apparatus comprising: An image processing method for correcting image inclination,
An image input step for inputting image data;
A line segment extraction step of extracting a line segment group in the image;
An angle detection step of detecting an angle between the line segment and a predetermined reference direction;
An image rotation step for rotating the image based on the detection result;
An image processing method comprising: The line segment extraction step includes:
An edge extraction step of extracting an edge group consisting of edges of the image;
An edge region extraction step of extracting an edge region group by extracting pixels in the vicinity of the edge;
A run length extraction step for extracting a run length group consisting of run lengths of a predetermined length or more based on the edge region group, and
The image processing method according to claim 9, wherein in the run length group, the line group group is extracted by integrating run lengths having a predetermined connection relationship.   An image processing program for operating the image processing apparatus according to any one of claims 1 to 8, wherein the image processing program causes a computer to function as each of the above means.   A computer-readable recording medium on which the image processing program according to claim 11 is recorded.

Images