JP2004045128A - Calibration reference device, image processing method for calibrating image inspection device using the calibration reference device, and image inspection device calibration method - Google Patents

Abstract

A image processing with high accuracy for the calibration of the image inspection device, can be executed at high speed, calibration reference device, and provides the calibrated reference device and image processing method using, and image inspection device calibration method.
A calibration reference has a planar shape, and two projections or dents having vertices are formed on a peripheral portion of the planar shape, and projections or dents formed at two locations are provided. Is used as a length measurement position for calibration of the image inspection apparatus. Then, in the image inspection apparatus, (1) the calibration reference is extracted on the image from the image data of the calibration reference, and (2) the projections or dents formed at two places are further displayed on the image from the extracted image. And (3) calculate the length between the vertices with pixel accuracy or sub-pixel accuracy, and (4) obtain an index value indicating the machine difference of the image inspection apparatus.
[Selection] Figure 2

Description

[Claims]
1. A calibration reference device for use in calibrating an image inspection apparatus, wherein the calibration reference device has a planar shape, and a projection or a dent having an apex is provided on a periphery of the planar shape. A calibration reference device formed at a position, and a distance between vertexes of the protrusions or the dents formed at the two positions is a length measurement position for calibration of the image inspection apparatus.
Wherein said calibration reference device is substantially circular in shape, the calibration standards as claimed in claim 1.
3. A projection having two vertices is formed at a position on a peripheral portion opposite to a center position of the substantially circular shape, and a distance between the vertices of the two projections is defined by the calibration reference. 3. The calibration reference of claim 2, which is the maximum diameter of the reference.
4. An image processing method for calibrating an image inspection apparatus, using the calibration reference device according to claim 1.
Extracting the calibration reference on an image from the image data of the calibration reference taken by the image inspection apparatus;
Extracting, on the image, the protrusions or dents formed at the two locations from the extracted image of the calibration reference device;
Calculating the length between the respective vertices of the extracted two protrusions or dents with pixel accuracy or sub-pixel accuracy;
Based on the calculated length between the vertices at the pixel accuracy or sub-pixel accuracy, and a known reference length obtained by measuring the distance between the vertices of the calibration reference device in advance, the image inspection device Obtaining an index value representing the machine difference;
An image processing method for calibrating an image inspection apparatus, comprising:
Wherein when extracting an image of said calibration datum from said image data, to perform the extraction using an automatic thresholding, for calibration of the image inspection apparatus according to claim 4, characterized in Image processing method.
6. An apparatus for calibrating machine differences of said image inspection apparatus based on an index value indicating machine differences of said image inspection apparatus obtained by the image processing method according to claim 4 or 5. Calibration method for image inspection equipment
7. Software for executing a method according to claim 4 on a computer.
DETAILED DESCRIPTION OF THE INVENTION
[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a method for calibrating an image inspection apparatus that inspects an object based on an image obtained by imaging the object.
[0002]
[Prior art]
2. Description of the Related Art An image inspection apparatus for inspecting an optical member such as a lens (hereinafter, referred to as an object) by image processing is known. In such an image inspection apparatus, first, a test object illuminated with light from a light source is photographed by a CCD camera or the like, and a predetermined process is performed on the photographed image to determine a defect factor present in the test object. Extract and classify by shape. Generally, defective factors include dust, scratches, fluff (filaments), and dirt. Further, depending on the image inspection apparatus, the test object is divided into a plurality of areas, and the causes of failure are classified into areas corresponding to the positions where the causes of failure occur. When the extracted failure factors are classified according to shape and area, the image inspection device quantifies each failure factor into a numerical value (quality score). Then, the quality of the test object is determined based on the total score of each defect factor.
[0003]
On the other hand, in such an image inspection apparatus, it is important to calibrate a difference between the image inspection apparatuses. Therefore, normally, the diameter of the calibration reference device 10 is obtained on an image by using the circular calibration reference device 10 having a known (measured) diameter as shown in FIG. First, the calibration reference device 10 is set in an image inspection apparatus, an image is taken, and an image is generated using a fixed threshold. From the image of the calibration reference device 10, obtains the diameter of the calibration reference device 10 as the number of pixels. Since the diameter of the calibration reference 10 is known, the “length per pixel” is obtained, for example, in (microns / pixel) for each image inspection apparatus. This value is an index indicating the difference between the image inspection apparatuses. Therefore, if this value (“length per pixel”) is obtained for each device, the image inspection device can be calibrated based on this value.
[0004]
[Problems to be solved by the invention]
FIGS. 10 and 11 show examples in which an image of the calibration reference device 10 is generated by two image inspection apparatuses having machine differences. FIG. 10 shows an image generated by the first image inspection apparatus. 10, the diameter of the first calibration standard 10 on the image generated by the image inspecting apparatus, which is larger Ku than the dimension of the way, the actual calibration reference 10 shaded circle 13 The state is shown. FIG. 11 shows an image generated by the second image inspection apparatus. FIG. 11 shows a state in which the diameter of the calibration reference device 10 is smaller than the actual size of the calibration reference device 10 as indicated by a hatched circle 15 on the image generated by the second image inspection apparatus. Is shown.
[0005]
In this case, based on the “length per pixel” determined for the first image inspection device and the second image inspection device, eliminating the machine difference between these image inspection devices, Can be calibrated .
[0006]
However, acquiring instrumental differences using a circular calibration reference means that the region of interest on the generated image is the entire image, that is, the diameter of the entire image is used to determine the diameter of the calibration reference. Partial decisions must be made. Further, since the light amount of the image inspection apparatus is not constant for each apparatus, the image of the calibration reference device may not be constant between apparatuses in a measurement using a fixed threshold as in the related art.
[0007]
In addition, although a circular calibration standard generally used in the past is used by measuring the diameter precisely in advance, it is usually difficult to make the calibration standard a strict circle, In addition, it is usually difficult to match the position where the diameter is determined by image processing with the position where the measurement is actually performed. Therefore, the position where the diameter is obtained on the image of the calibration reference device may be different from the position where the actual measurement is performed.
[0008]
The present invention has been made in view of the above circumstances. That is, the present invention solves the above problems, with high accuracy, high-speed image processing for calibration of the image inspection device, a calibration reference device, an image processing method using the calibration reference device, And a method of calibrating an image inspection apparatus.
[0009]
[Means for Solving the Problems]
In order to achieve the above object, the calibration reference device has a planar shape, and two projections or depressions having vertices are formed on the peripheral edge of the planar shape, and the projections formed at two locations are formed. Alternatively, the configuration is such that the space between the vertices of the recess is used as a length measurement position for calibration of the image inspection apparatus. The vertices of the two protrusions or dents serve as markers for performing physical dimension measurement and indicating the place where the measurement was performed, and at the same time, the protrusions or dents themselves are displayed on the image. This is a mark for extracting the point for which the length is to be calculated. Therefore, if this calibration reference device is used, the position where the length should be calculated can be quickly recognized on the image, and the image processing can be performed efficiently.
[0010]
In this case, the calibration reference device may have a substantially circular shape (claim 2). Further, the calibration reference device has two protrusions having vertexes formed at positions on a peripheral portion opposed to the center position of the substantially circular shape, and a distance between the vertices of the two protrusions is set to the calibration reference device. May be the maximum diameter (claim 3).
[0011]
To achieve the above object, using a calibration reference having a protrusion or a dent as described above, as a method for calibrating an image inspection apparatus,
(1) extracting a calibration reference on an image from image data of the calibration reference taken by the image inspection apparatus;
(2) extracting, on the image, protrusions or dents formed at two locations from the extracted image of the calibration reference device;
(3) calculating the length between the vertices of each of the two extracted protrusions or dents with pixel accuracy or sub-pixel accuracy;
(4) Based on the calculated length between vertices at the pixel accuracy or sub-pixel accuracy and the known reference length obtained by measuring the distance between the vertices of the calibration reference device in advance, Obtaining an index value representing the difference. Since the calibration reference device has a protrusion or a depression as a mark, it is possible to quickly extract a position on the image at which the length is to be calculated.
[0012]
In the step (1), when the image of the calibration reference device is extracted from the image data, the image may be extracted using automatic threshold processing.
[0013]
A calibration method for calibrating the machine difference of the image inspection device can be configured based on the index value indicating the machine difference of the image inspection device obtained by the image processing method as described above (claim 6).
[0014]
BEST MODE FOR CARRYING OUT THE INVENTION
FIG. 1 is a schematic configuration diagram of the image inspection apparatus 1. Image processing for machine difference calibration as an embodiment of the present invention is performed on the image inspection apparatus 1. First, the configuration and normal operation of the image inspection apparatus 1 will be described with reference to FIG.
[0015]
The image inspection apparatus 1 includes an image photographing unit 100 and an image processing system 200. The image photographing unit 100 includes a CCD camera 101, a light source 102, and a holder 104. The upper surface of the holder 104 is configured as a table 104a on which the test object S is placed. The lower portion of the holder 104 is provided on a rail 103a formed on the upper surface of the stage 103 of the image capturing unit 100, and is configured as a driving unit 104b that can move in the horizontal direction along the rail 103a. Further, the CCD camera 101 and the light source 102 are fixed to the frame unit 105 of the image capturing unit 100, respectively. The CCD camera 101 has an objective lens unit 101a and a CCD 101b. The light source 102 has a plurality of optical fibers. Each optical fiber is disposed such that light emitted from each of the optical fibers is obliquely incident on the test object S and the emission end faces are arranged in a ring.
[0016]
The image processing system 200 processes the image data transmitted from the image photographing unit 100 to determine the quality of the test object S, and the image data transmitted from the image photographing unit 100 and image processing by the processor 201. A monitor 202 capable of displaying results and the like.
[0017]
Next, a process for calibrating machine differences of the image inspection apparatus 1 will be described. FIG. 2 shows a calibration reference 21 used for performing calibration . FIG. 3 is a flowchart of a calibration process using the calibration reference device 21, and FIG. 4 is a detailed flowchart of the image processing in FIG.
[0018]
The calibration reference 21 (FIG. 2) has a substantially circular shape, has two projections 21a and 21b, and is configured so that the maximum diameter D is between the projections 21a and 21b. Therefore, the measurement of the dimensions of the calibration reference device 21 may be performed between the protrusions 21a and 21b. In the image processing, the maximum diameter D can be obtained as the number of pixels (or with sub-pixel accuracy) by focusing only on the protrusions 21a and 21b of the calibration reference device 21.
[0019]
As shown in FIG. 3, the flow of the machine difference calibration is as follows. First, the calibration reference device 21 is installed in the image inspection apparatus 1, and the image capturing unit 100 performs scanning to capture an image of the calibration reference device 21 (S301). The multi-gradation image data obtained by the imaging is transmitted to the image processing system 200, where the image processing is performed (S302). In the image processing (S302), the maximum diameter D of the calibration reference device 21 is calculated, whereby the "length per pixel" in the image inspection apparatus 1 is obtained, and based on this value, the image processing apparatus 1 is calibrated (S303).
[0020]
FIG. 4 shows details of the image processing (S302) of FIG. First, automatic threshold processing is performed on a multi-tone image obtained by imaging, and the calibration reference device 21 is extracted on the image (S401). As the automatic threshold processing, a method of finding a suitable threshold from a histogram of an image can be used. The accuracy of extracting the calibration reference device 21 can be improved by the automatic threshold processing. The extracted image is as shown in FIG.
[0021]
Next, in step S402, a region of interest is narrowed down from the extracted image (FIG. 5) of the calibration reference device 21. It is relatively easy to narrow down the region of interest because the projections 21a and 21b may be found on the image. The region of interest is determined as a region including the protruding portions 21a and 21b, as shown as regions 51 and 52 in FIG. The processing up to this point is performed on a pixel-by-pixel basis.
[0022]
Next, in step S403, the boundary position of the calibration reference device 21 is determined for the regions of interest 51 and 52 with sub-pixel accuracy by sub-pixel processing. Sub-pixel processing refers to, for example, determining a luminance gradient near a boundary (that is, an approximate waveform of the gradient) based on gradation data of each pixel, and applying a predetermined threshold to the luminance gradient to determine a boundary position. This is the processing to be performed. By the sub-pixel processing, for example, the boundary position can be determined with 0.1 pixel accuracy. FIG. 7 shows an example of a boundary position determined with sub-pixel accuracy with respect to the region of interest 52 (boundary position B1).
[0023]
Since the boundary regions have been determined for the regions of interest 51 and 52, the maximum distance between the regions of interest 51 and 52 (that is, the length between the vertices of the regions of interest 51 and 52) is calculated in step S404. As described above, in the image processing of FIG. 4, the processing may be performed while focusing on the regions of interest 51 and 52, and thus the processing speed may be increased. Therefore, even if image processing with a relatively heavy load such as sub-pixel processing is employed, the overall processing speed can be increased.
[0024]
In addition, while monitoring the maximum distance calculated by the series of steps in FIG. 4 (repeating the series of steps in FIG. 4), control for adjusting the focus of the CCD camera 101 is performed. It is preferable that the focus adjustment is completed when the distance becomes minimum, and that the distance at this time is determined as a final maximum distance. The maximum distance thus obtained is used in the next calibration processing (S303).
[0025]
In step S303 ( calibration processing), the magnification (1) of the image inspection apparatus 1 is determined based on the maximum distance obtained as described above and the actual dimensions (measured dimensions) of the maximum diameter portion of the calibration reference device 21. The length per pixel (microns / pixel) is determined and stored. This magnification serves as an index value indicating a machine difference of the image inspection apparatus. This magnification is used as a correction value for correcting the magnitude of the defect factor obtained by the image processing during the actual inspection of the test object. Thereby, the machine difference between the image inspection apparatuses is calibrated .
[0026]
For example, it is assumed that magnifications are obtained and stored as P1 (microns / pixel) and P2 (microns / pixel) in two image inspection apparatuses, respectively. Then, it is assumed that a defect factor having a length of 2 pixels is detected in each image inspection apparatus. In this case, in each image inspection apparatus, the size of the defect factor is determined as 2 · P1 (micron) and 2 · P2 (micron). Since the detected number of pixels is not used as the dimension of the cause of the defect as it is, but a value corrected by multiplying each image inspection apparatus by a magnification specific to the image inspection apparatus is used as the dimension of the cause of the defect, a correct dimension with machine differences eliminated is obtained.
[0027]
There can be various modifications of the embodiments of the present invention described above. For example, in the image processing shown in FIG. 4, the automatic threshold processing (S401) is effective to enhance the accuracy of the extraction of the calibration reference device, but the automatic threshold processing is not always necessary, and the normal empirical processing is performed. A configuration using a threshold may be used. Also, the sub-pixel processing (S403) is not necessarily required, and may be performed in a normal pixel unit.
[0028]
Further, as can be understood from the above, the calibration reference device is not limited to the configuration shown in FIG. As long as there is a mark of a part whose size has been measured and it is relatively easy to extract that part as a region of interest on an image, a calibration standard having various shapes can be used. Further, the portion where the dimension is measured does not necessarily have to be the maximum diameter of the calibration reference device.
[0029]
Figure 8 shows another example of a calibration reference device (the calibration reference 60). As shown in FIG. 8, the calibration reference unit 60 is formed with dents 61 and 62. Since it is relatively easy to extract the dents 61 and 62 as the region of interest by image processing, The calibration reference 60 can achieve the same function as the calibration reference 21. In this case, the length D2 between the vertices of the depressions 61 and 62 is measured in advance, and the length corresponding to D2 is calculated in image processing.
[0030]
【The invention's effect】
As described above, according to the calibration reference device of the present invention, the position where the length is to be calculated can be quickly recognized on the image, and the image processing can be performed efficiently.
[Brief description of the drawings]
FIG. 1 is a diagram illustrating a configuration of an image inspection apparatus in which an image processing method according to an embodiment of the present invention is executed.
FIG. 2 is a diagram illustrating a calibration reference used in the image processing of FIG. 1;
FIG. 3 is a flowchart illustrating a calibration process for calibrating the image inspection apparatus of FIG. 1;
FIG. 4 is a flowchart illustrating details of image processing in the calibration processing of FIG. 3;
FIG. 5 is a diagram showing an image of a calibration reference device extracted in the image processing of FIG. 4;
FIG. 6 is a diagram showing a region of interest extracted in the image processing of FIG. 4;
FIG. 7
FIG. 7 is a diagram showing a state where sub-pixel processing has been performed on the region of interest in FIG. 6.
FIG. 8 is a diagram showing another example as a calibration reference device.
FIG. 9
FIG. 10 is a diagram illustrating a conventional calibration reference device.
FIG. 10 is a diagram showing a state in which the calibration reference device is extracted larger than the actual condition.
FIG. 11 is a diagram showing a state in which the calibration reference device is extracted smaller than the actual one.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 Image inspection apparatus 21 Calibration standard 100 Image photographing part 101 CCD camera 102 Light source 104 Holder 104a Table part 200 Image processing system 201 Processor 202 Monitor

Claims (7)

A calibration reference device for use in calibrating an image inspection device, wherein the calibration reference device has a planar shape, and a peripheral portion of the planar shape is formed with two protrusions or dents having vertices, A calibration standard, wherein a distance between vertices of the protrusions or dents formed at two locations is a length measurement position for calibration of the image inspection apparatus. The calibration standard of claim 1, wherein the calibration standard has a substantially circular shape. Two protrusions having vertices are formed at positions on the peripheral edge portion opposite to the center position of the substantially circular shape, and the distance between the vertices of the two protrusions is the maximum diameter of the calibration reference device. 3. The calibration reference of claim 2, wherein: An image processing method for calibrating an image inspection device, using the calibration reference device according to claim 1,
Extracting the calibration reference on an image from image data of the calibration reference taken by the image inspection device;
Extracting, on the image, protrusions or dents formed at the two locations from the extracted image of the calibration reference device;
Calculating the length between the respective vertices of the extracted two protrusions or dents with pixel accuracy or sub-pixel accuracy;
Based on the calculated length between the vertices at the pixel accuracy or sub-pixel accuracy, and a known reference length obtained by measuring the distance between the vertices of the calibration reference device in advance, the image inspection device of the image inspection device Obtaining an index value representing the machine difference;
An image processing method for calibrating an image inspection apparatus, comprising: 5. The image processing method for calibrating an image inspection apparatus according to claim 4, wherein when extracting an image of the calibration reference device from the image data, the extraction is performed using automatic threshold processing. An image inspection apparatus according to claim 4, wherein the difference between the image inspection apparatuses is calibrated based on an index value indicating the difference between the image inspection apparatuses obtained by the image processing method according to claim 4. Calibration method. Software for executing the method according to any one of claims 4 to 6 on a computer.

Images