JP2009294170A - Apparatus, method, and program for defect detection and recording medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a defect detection apparatus and a defect detection method capable of shortening inspection time without sacrificing detection accuracy. <P>SOLUTION: The defect detection apparatus 100 includes: a high-resolution image construction processing part 123 for generating high-resolution images by referencing to a plurality of pick-up images acquired by imaging an object to be evaluated 200; a defect detection processing part 124 for detecting defects in the object to be evaluated 200 by making reference to the high-resolution images; and an optical system control part 122 for setting the number of pick-up images which the high-resolution image construction processing part 123 referred for generating the high-resolution images based on the defect shape information recorded in the recording part 121. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a defect detection apparatus, a defect detection method, and a defect detection program for detecting a defect in an evaluation object based on an image obtained by imaging the evaluation object. The present invention also relates to a recording medium on which such a defect detection program is recorded.

  In recent years, display devices such as televisions and monitors have been made thinner and larger, and their demand has increased. Accordingly, display devices are required to have higher display quality than ever before. In manufacturing a display device, it is important to accurately detect defects such as bright spots and black spots.

  As a technique for detecting defects such as bright spots and black spots in a display device, a technique in which an investigator visually detects the defects has been conventionally used. However, the visual detection causes problems such as oversight due to fatigue of the investigator and variation in evaluation by the investigator. Therefore, a technique for automatically detecting a defect using a captured image obtained by imaging a display device has been introduced.

  In order to detect a defect accurately using a captured image, it is preferable to use a high-resolution captured image. However, in order to obtain a high-resolution captured image, it is usually necessary to image the display device using a large and expensive high-resolution camera, resulting in space problems and cost problems. Therefore, it is conceivable to detect defects with high accuracy without using a high-resolution camera by generating a high-resolution image having a higher resolution than the captured image from a plurality of captured images.

  As an example of a resolution enhancement process for generating a resolution-enhanced image from a plurality of captured images, a pixel shifting method can be given. The pixel shifting method is, for example, combining a first captured image obtained by imaging a certain subject and a second captured image obtained by shifting the subject by a half pixel pitch. This is a process for generating a high-resolution image.

As a technique for improving the efficiency of such high resolution processing, for example, those described in Patent Documents 1 and 2 are known. Japanese Patent Application Laid-Open No. 2004-151561 describes a technique for increasing the efficiency of calculation of a positional deviation amount between images and high resolution processing by separating the images into bands. Patent Document 2 describes a technique for preventing unsuccessful shooting of pixel shift shooting due to a change in performance of a pixel shift mechanism or the like.
Japanese Patent Laying-Open No. 2005-352720 (published on December 22, 2005) JP 2002-218328 A (released on August 2, 2002)

  When applying high resolution processing such as pixel shifting to defect detection, it is possible to avoid space problems and cost problems associated with the use of high resolution cameras, but to image the evaluation object multiple times. There is a problem that the inspection time increases due to the imaging time required and the processing time required to synthesize a plurality of captured images. In particular, it is difficult to reduce the imaging time required for imaging the evaluation object a plurality of times even if the techniques described in Patent Documents 1 and 2 are used to improve the efficiency of the high resolution processing. It was.

  The present invention has been made in view of the above problems, and an object of the present invention is to realize a defect detection apparatus and a defect detection method that reduce the inspection time without sacrificing detection accuracy. is there.

  In order to solve the above-described problem, the defect detection apparatus according to the present invention generates a high-resolution image that has a higher resolution than the resolution of the captured image by referring to a plurality of captured images obtained by imaging the evaluation object. Based on the defect shape information indicating the shape of the defect to be detected, the detection means for detecting the defect in the evaluation object with reference to the generation means for generating, the high-resolution image generated by the generation means Setting means for setting the number of the picked-up images referred to by the generating means for generating a resolution image.

  In high resolution processing that generates a high resolution image with reference to a plurality of captured images, the number of captured images required to detect a defect with a predetermined detection accuracy depends on the shape of the defect to be detected. Different.

  According to said structure, the number of the captured images referred in order to produce | generate a high resolution image can be set based on the defect shape information which shows the shape of the defect which should be detected. Therefore, it is possible to reduce the number of captured images referred to in order to generate a high-resolution image while maintaining a predetermined detection accuracy. For this reason, there exists an effect that inspection time can be shortened, maintaining predetermined detection accuracy.

  In the defect detection apparatus according to the present invention, it is preferable that the generation unit generates the high-resolution image by a pixel shifting method.

  According to the pixel shifting method, the short axis direction and / or the long axis are obtained by referring to two captured images obtained by shifting the image of the evaluation object in the short axis direction of the defect to be detected. A resolution-enhanced image that can detect a defect with the same detection accuracy as that obtained when four captured images obtained by shifting the evaluation object in the direction and capturing images is generated.

  For this reason, according to said structure, test | inspection time can be shortened, maintaining a predetermined detection accuracy.

  The defect detection apparatus according to the present invention further includes a recording unit that records defect shape information indicating the shape of the defect estimated in advance, and the setting unit is based on the defect shape information recorded in the recording unit. Thus, it is preferable to set the number of the captured images referred to by the generation unit in order to generate the resolution-enhanced image.

  According to said structure, inspection time can be shortened, maintaining a predetermined detection precision with respect to the evaluation object which can estimate the shape of the defect which should be detected previously.

  The defect detection apparatus according to the present invention further includes a recording unit that records defect shape information indicating a shape of the defect detected by the defect detection unit, and the setting unit records the defect recorded in the recording unit. It is preferable that the number of the captured images referred to by the generating unit to set the resolution-enhanced image is set based on the shape information.

  According to the above configuration, predetermined detection accuracy is maintained by repeatedly detecting defects in the same or similar evaluation object even for an evaluation object in which the shape of the defect to be detected cannot be estimated in advance. However, the inspection time can be shortened.

  In the defect detection apparatus according to the present invention, the setting means detects in which direction the evaluation object is shifted and picked up to generate a high-resolution image with reference to a picked-up image. It is preferable that the setting is based on the aspect ratio of the power defect.

  According to the above configuration, in which direction the defect to be detected is detected in which direction the high-resolution image is generated with reference to the captured image obtained by shifting the image of the evaluation object. It can be set depending on whether the defect is a stretched defect.

  In the defect detection apparatus according to the present invention, the setting means includes two captured images obtained by shifting the image of the evaluation object in the first direction and capturing the evaluation object in the first direction. Two captured images obtained by shifting images in the vertical second direction, or acquired by shifting the evaluation object in the first direction and / or the second direction. It is preferable to set which of the four captured images to be used to generate the high-resolution image.

  According to said structure, when it is a defect which makes a 1st direction a short axis, refer to the two captured images obtained by image-shifting an evaluation target object in a 1st direction A high resolution image can be generated. When the defect has a minor axis in the second direction, the high-resolution image is obtained by referring to two captured images obtained by imaging the evaluation object by shifting in the second direction. Can be generated. In addition, when neither of them is obtained, a high-resolution image is obtained by referring to two captured images obtained by capturing the evaluation object by shifting in the first direction and / or the second direction. Can be generated. Thereby, in any case, it is possible to detect a defect with the same detection accuracy as when four captured images are referred to.

  In the defect detection apparatus according to the present invention, the detection means calculates a contrast of each pixel constituting the high-resolution image, and sets a set of pixels in which the calculated contrast exceeds a predetermined threshold as a defect image. It is preferable to specify.

  According to said structure, when an evaluation target object is a uniform flat surface of a single color, the different color part which exists in an evaluation target object can be detected as a defect.

  In the defect detection apparatus according to the present invention, the detection means may use a reference pixel determined according to the pixel value of the pixel and the pattern of the evaluation object for the contrast of each pixel constituting the high-resolution image. It is preferable that the calculation is performed by taking the absolute value of the difference from the average pixel value of the reference pixels estimated to have the same pixel value as that pixel when there is no defect.

  According to said structure, a defect can be detected also with respect to the multicolor evaluation target object which has a periodic pattern. For example, when the evaluation object is a liquid crystal display panel composed of pixels having three colors of RGB, the presence or absence of defects in the R color portion of a pixel corresponds to the R color portion of each pixel adjacent to the pixel in the upper, lower, left, and right directions. Therefore, it is possible to accurately determine the contrast by using the pixels in the high resolution image as a reference image.

  In the defect detection apparatus according to the present invention, it is preferable that the detection means calculates a defect intensity by integrating the contrast of each pixel constituting the defect image.

  According to said structure, the intensity | strength of the detected defect can be calculated easily.

  In order to solve the above-described problem, a defect detection method according to the present invention refers to a plurality of captured images obtained by imaging an evaluation object with an imaging device, and achieves a higher resolution than that of the imaging device. Based on a generation step for generating an image, a detection step for detecting a defect in the evaluation object with reference to the high-resolution image generated in the generation step, and defect shape information indicating the shape of the defect to be detected And a setting step of setting the number of captured images referred to in the generation step in order to generate the high resolution image.

  In high resolution processing that generates a high resolution image with reference to a plurality of captured images, the number of captured images required to detect a defect with a predetermined detection accuracy depends on the shape of the defect to be detected. Different.

  According to said structure, the number of the captured images referred in order to produce | generate a high resolution image can be set based on the defect shape information which shows the shape of the defect which should be detected. Therefore, it is possible to reduce the number of captured images referred to in order to generate a high-resolution image while maintaining a predetermined detection accuracy. For this reason, there exists an effect that inspection time can be shortened, maintaining predetermined detection accuracy.

  The defect detection apparatus according to the present invention may be realized by a computer. In this case, a defect detection program for realizing the defect detection apparatus in the computer by operating the computer as each of the above means and a computer-readable recording medium on which the defect detection program is recorded also fall within the scope of the present invention. .

  The defect detection apparatus according to the present invention refers to a plurality of captured images obtained by imaging an evaluation object, and generates a resolution-enhanced image having a resolution higher than that of the captured image, and the generation unit In order to generate the resolution-enhanced image based on the detection means for detecting the defect in the evaluation object and the defect shape information indicating the shape of the defect to be detected with reference to the resolution-enhanced image generated by Setting means for setting the number of captured images referred to by the generating means.

  In addition, the defect detection method according to the present invention refers to a plurality of captured images obtained by imaging an evaluation object with an imaging device, and generates a resolution-enhanced image having a resolution higher than that of the imaging device. And the resolution enhancement based on the detection step of detecting the defect in the evaluation object with reference to the resolution-enhanced image generated in the generation step and the defect shape information indicating the shape of the defect to be detected. A setting step for setting the number of captured images referred to in the generation step in order to generate an image.

  Therefore, it is possible to reduce the number of captured images referred to in order to generate a high-resolution image while maintaining a predetermined detection accuracy. For this reason, there exists an effect that inspection time can be shortened, maintaining predetermined detection accuracy.

  The defect detection apparatus according to this embodiment will be described below with reference to the drawings.

(Configuration of defect detection device)
First, the defect detection apparatus 100 according to the present embodiment will be described with reference to FIG. FIG. 1 is a block diagram showing a main configuration of the defect detection apparatus 100.

  As shown in FIG. 1, the defect detection apparatus 100 is an apparatus that detects a defect in the evaluation object 200 and includes an optical system 110 and a PC 120. As the evaluation object 200, a display device such as a liquid crystal display panel is assumed.

  The optical system 110 includes a camera 111, a camera driving device 112, a stage 113, and a stage driving device 114. After moving the stage 113 on which the evaluation object 200 is placed by the stage driving device 114, the camera 111 can release the shutter of the camera 111 so that the evaluation object 200 can be imaged from any position. It is configured. Here, the relative position between the evaluation object 200 and the camera 111 is moved by moving the stage 113 on which the evaluation object 200 is placed. However, by moving the camera 111, the evaluation object is moved. The relative position between the object 200 and the camera 111 may be moved.

  The stage driving device 114 can move the stage 113 in the first movement direction and the second movement direction perpendicular to the first movement direction. When the evaluation object 200 is a liquid crystal display panel, the first direction is selected to match the long side direction of the liquid crystal display panel, and the second direction is set to match the short side direction of the liquid crystal display panel. Chosen. In the following description, “vertical” or “vertical direction” refers to the first direction, and “horizontal” or “horizontal direction” refers to the second direction.

  The PC 120 includes a recording unit 121, an optical system control unit 122, a high resolution processing unit 123, and a defect detection processing unit 124. The optical system control unit 122 performs an imaging position setting process. In other words, the camera driving device 112 sets the position from which the evaluation object 200 is to be imaged based on the defect information recorded in the recording unit 121, and gives instructions necessary to image the evaluation object 200 from that position. And the stage driving device 114. The imaging position setting process executed by the optical system control unit 122 will be described in detail later with reference to another drawing.

  The high resolution processing unit 123 executes high resolution processing. That is, a high resolution image is generated from an image captured by the camera 111. The defect detection processing unit 124 executes defect detection processing. That is, a defect is detected from the high resolution image generated by the high resolution processing unit 123. The high resolution processing executed by the high resolution processing unit 123 and the defect detection processing executed by the defect detection unit 124 will be described in detail later with reference to different drawings.

(Operation of defect detection device)
Next, an operation of the defect detection apparatus 100, that is, a defect detection method using the defect detection apparatus 100 will be described with reference to FIG. 2 and FIG.

  FIG. 2 is a flowchart showing a first defect detection method suitable when the size and shape of the defect to be detected are known. In this defect detection method, defect information indicating a known defect size and defect shape is recorded in the recording unit 121, and defects in the evaluation object 200 are efficiently detected by referring to the defect information. As the defect information indicating the defect shape, for example, an aspect ratio can be used.

  The outline of each process included in the first defect detection method shown in FIG. 2 will be described as follows.

  That is, in step S11, based on the defect information read from the recording unit 121, the optical system control unit 122 sets a position from which the evaluation object 200 is imaged, and images the inspection object 200 from the position. Instructions necessary for this are given to the camera driving device 112 and the stage driving device 114.

  In step S12, the camera drive unit 112 and the stage drive unit 114 drive the camera 111 and the stage 113 based on an instruction from the optical system control unit 122, so that the object to be evaluated is set from the position set in step S11. 200 is imaged.

  When the defect size is larger than a preset threshold value, one image is taken in step S12 to obtain one image. If not, imaging is performed twice or four times in step S12, and two or four images are obtained.

  In step S13, the high resolution processing unit 123 executes the high resolution processing to obtain a high resolution image from the image obtained in step S12. However, when the defect size is larger than a preset threshold value, the high resolution processing is not performed.

  In step S14, the defect detection processing unit 124 detects a defect in the evaluation object 200 from the high-resolution image obtained in step S13 by executing the defect detection process. However, when the defect size is larger than a preset threshold value, the defect is detected from the captured image itself obtained in step S12.

  Examples of the case where the size and shape of the defect to be detected are known in advance include a case where a black spot defect or a bright spot defect in a liquid crystal display panel is detected. In this case, the size and shape of the defect to be detected can be predicted in advance by specifying the pixel size and pixel shape of the liquid crystal display panel from the model information of the liquid crystal display panel that is the evaluation object. Therefore, defect information indicating the size and shape of the defect to be detected can be recorded in the recording unit 121 in advance.

  FIG. 3 is a flowchart showing a second defect detection method suitable when the size and shape of the defect to be detected are unknown. In this defect detection method, defect information indicating the defect size and defect shape detected by the defect detection processing unit 124 is recorded in the recording unit 121, and the defect in the evaluation object 200 is detected by referring to the defect information. Detect efficiently. As the defect information indicating the defect shape, an aspect ratio or the like can be used as in the first defect detection method.

  The outline of each process included in the second defect detection method shown in FIG. 3 will be described as follows.

  That is, in step S <b> 21, the optical system control unit 122 sets from which position the evaluation object 200 is imaged based on the defect information read from the recording unit 121, and images the evaluation object 200 from that position. Instructions necessary for this are given to the camera driving device 112 and the stage driving device 114. However, when performing step S21 for the first time, since defect information has not yet been recorded in the recording unit 121, an appropriate imaging position is set so that a high-resolution image can be obtained to some extent by the high-resolution processing. .

  In step S22, the camera drive unit 112 and the stage drive unit 114 drive the camera 111 and the stage 113 based on an instruction from the optical system control unit 122, so that the object to be evaluated is set from the position set in step S21. 200 is imaged.

  When the defect size is larger than a preset threshold value, one image is taken in step S12 to obtain one image. If not, imaging is performed twice or four times in step S12, and two or four images are obtained.

  In step S23, the high resolution processing unit 123 executes the high resolution processing, thereby obtaining a high resolution image from the image in step S22. However, when the defect size is larger than a preset threshold value, the high resolution processing is not performed.

  In step S24, the defect detection processing unit 124 executes defect detection processing to detect defects in the evaluation object 200 from the high-resolution image obtained in step S23, and to detect the size and size of the detected defects. The defect information indicating the shape is written in the recording unit 121. When the defect size is larger than a preset threshold value, the defect is detected from the captured image itself obtained in step S22, and defect information indicating the size and shape of the detected defect is written in the recording unit 121. Thereafter, the defect information written in the recording unit 121 in step S24 is referred to in step S21.

  According to the second defect detection method shown in FIG. 3, the defect information indicating the size and shape of the defect to be detected is based on the size and shape of the defect previously detected in the same or similar evaluation object. Can be determined. For this reason, the resolution of the high-resolution image obtained by the high-resolution processing can be suppressed to a necessary minimum value. In other words, the number of images used for the resolution enhancement processing can be suppressed to the minimum necessary number. Therefore, an unnecessary increase in inspection time can be avoided.

  Further, as will be described later, since the position where the evaluation object 200 is imaged is determined by referring not only to the size of the defect but also to the shape of the defect (for example, the aspect ratio), Only the resolution in the vertical direction can be increased, and only the resolution in the horizontal direction can be increased for vertically long defects. That is, by setting only the resolution in the minor axis direction of the defect high, it is possible to reduce the number of images used for the high resolution processing while ensuring the resolution necessary for the defect detection processing.

(Imaging position setting process)
Next, the imaging position setting process executed by the optical system control unit 122 in the above-described steps S11 and S21 will be described with reference to FIG. FIG. 4 is a flowchart illustrating an example of the imaging position setting process.

  The optical system control unit 122 first compares the defect size L indicated by the defect information with a predetermined threshold Th0 (S31). When the defect size L is larger than the threshold value Th0, the optical system control unit 122 sets only a predetermined reference position for imaging the evaluation object 200 as an imaging position (S32). In this case, in steps S12 and S22, one image is obtained by one imaging.

  When the defect size L is smaller than the threshold Th0, the optical system control unit 122 sets the aspect ratio R (defect vertical width / defect horizontal width) indicated by the defect information to predetermined thresholds Th1 and Th2 (Th1> 1). > Th2) (S33).

  When the aspect ratio R is larger than the threshold value Th1, that is, when the defect is vertically long, the optical system control unit 122, in addition to the reference position, shifts a position shifted by a half pixel pitch horizontally with respect to the reference position. (S34). At this time, two images in which the position of the subject is laterally shifted by a half pixel pitch are obtained in steps S12 and S22.

  When the aspect ratio R is smaller than the threshold Th2, that is, when the defect is horizontally long, the optical system control unit 122, in addition to the reference position, captures a position shifted by a half pixel pitch vertically with respect to the reference position. (S35). At this time, two images in which the position of the subject is vertically shifted by a half pixel pitch are obtained in steps S12 and S22.

  When the aspect ratio R is greater than or equal to the threshold Th2 and less than or equal to the threshold Th1, the optical system control unit 122 sets the reference position at a position shifted by a half pixel pitch horizontally with respect to the reference position. On the other hand, a position shifted by a half pixel pitch vertically and a position shifted by a half pixel pitch horizontally and vertically from the reference position are set as imaging positions (S36). At this time, four images in which the position of the image of the evaluation object 200 in the image is shifted by a half pixel pitch vertically and horizontally are obtained in steps S12 and S22.

  Note that the imaging position setting process shown in FIG. 4 is merely an example of a possible imaging position setting process, and various modifications are possible.

(High resolution processing)
Next, the high resolution processing executed by the high resolution processing unit 123 in steps S13 and S23 described above will be described with reference to FIGS. The high resolution processing unit 123 generates a high resolution image using any one of the 2 × 1 pixel shift method, the 1 × 2 pixel shift method, and the 2 × 2 pixel shift method.

  FIG. 5 is an explanatory diagram showing an outline of the high resolution processing using the 2 × 1 pixel shifting method.

  As shown in FIG. 5A, the 2 × 1 pixel shift method is an image 501 obtained by imaging the evaluation object 200 from the reference position, and a half pixel pitch (each of the horizontal positions from the reference position). A high-resolution image 503 having twice the horizontal resolution is generated by combining the image 502 obtained by imaging the evaluation object 200 from the position where each square of the image corresponds to the pixel). Is the method.

  FIG. 5A conceptually shows from which position an image obtained by imaging is used for the high resolution processing. The actual image 501 and the image 502 are shown in FIG. ) As shown. The synthesis of the image 501 and the image 502 can be realized, for example, by calculating an average value of pixel values.

  The 1 × 2 pixel shift method is obtained by imaging the evaluation object 200 from an image obtained by imaging the evaluation object 200 from the reference position and a position moved by a half pixel pitch vertically from the reference position. This is a method of generating a high-resolution image having double vertical resolution by synthesizing with an image to be generated.

  FIG. 6 is an explanatory diagram showing an outline of the high resolution processing using the 2 × 2 pixel shifting method. As in FIG. 5A, FIG. 6 conceptually shows from which position an image obtained by imaging is used for the high resolution processing, and does not show an actual image. .

  As shown in FIG. 6, the 2 × 2 pixel shift method is based on an image 601 obtained by imaging the evaluation object 200 from the reference position and a position moved laterally by a half pixel pitch from the reference position. An image 602 obtained by imaging the evaluation object 200, an image 603 obtained by imaging the evaluation object 200 from a position moved by a half pixel pitch vertically from the reference position, and horizontal and vertical from the reference position. In this method, a high-resolution image 605 having twice the vertical and horizontal resolution is generated by combining the image 604 obtained by imaging the evaluation object 200 from the position moved by a half pixel pitch. is there.

(Defect detection processing)
Next, the defect detection process executed by the defect detection processing unit 124 in the above-described steps S14 and S24 will be described with reference to FIGS.

  The defect detection process is realized, for example, by calculating the contrast of each pixel constituting the high-resolution image and specifying a set of pixels in which the calculated contrast exceeds a predetermined threshold as a defect image. Furthermore, the strength of the defect may be evaluated by using a contrast volume calculated by integrating the contrast of the pixels included in the defect image.

  FIG. 7 is an explanatory diagram for explaining a contrast calculation method suitable for the case where the defect size corresponds to one pixel of the high-resolution image.

As shown in FIG. 7A, the contrast C of the pixel “a” is determined by the pixel value of the pixel “a” and the four pixels “b”, “c”, “d”, “e” adjacent to the pixel “a”. Can be calculated by taking the absolute value of the difference from the average pixel value of:
Contrast C = | a− (b + c + d + e) / 4 | (1).

Further, as shown in FIG. 7B, the pixel value of the pixel “a” and the eight pixels “b”, “c”, “d”, “e”, “f”, “g” located around the pixel “a”. It may be calculated by taking the absolute value of the difference from the average value of the pixel values of “h” and “i”:
Contrast C = | a− (b + c + d + e + f + g + h + i) / 8 | (2).

  When the evaluation target 200 has a pattern with a period corresponding to one pixel of the high-resolution image, refer to adjacent pixels and / or peripheral pixels as shown in FIGS. 7 (a) to 7 (b). When the contrast is calculated, the pixel value of the target pixel and the reference pixel are different from each other even if there is no defect.

  In such a case, if there is no defect, contrast may be calculated using a pixel estimated to have the same pixel value as the pixel “a” as a reference pixel. For example, as shown in FIG. 7C, when the pattern period of the evaluation object 200 corresponds to three pixels of the high-resolution image, two pixels are sandwiched as shown in FIG. Pixels “b”, “c”, “d”, “e” facing each side of the pixel “a”, and / or pixels “f”, “g”, “” that share a diagonal line with the pixel “a” across the two pixels. The contrast of the pixel “a” may be calculated using h and “i” as reference pixels.

  FIG. 8 is an explanatory diagram for explaining a contrast calculation method suitable for the case where the defect size corresponds to several pixels of the high resolution image. Here, it is assumed that the evaluation object 200 has a pattern with a period corresponding to 24 pixels of the high resolution image.

  The contrast C of the pixel 901 constituting the high-resolution image may be calculated by using a pixel estimated to have the same pixel value as the pixel 901 if there is no defect as a reference pixel. More specifically, as shown in FIG. 8A, four pixels facing each side of the pixel 901 across 23 pixels (in FIG. 8A, a reference pixel facing the left side of the pixel 901) (Only 902 is shown) and / or 4 pixels sharing a diagonal line with the pixel 901 across 23 pixels may be used as reference pixels and calculated according to the equation (1) or (2). As a result, as shown in FIG. 8B, a distribution of contrast is obtained in which the value increases only in the portion corresponding to the defect (in FIG. 8B, the high-contrast pixels are black and the contrast is high). Low pixels are shown in white).

(Effectiveness of 2 × 1 pixel shifting method and 1 × 2 pixel shifting method)
Finally, the effectiveness of the 2 × 1 pixel shifting method and the 1 × 2 pixel shifting method will be described with reference to FIGS. 9 and 10. Here, considering a vertically long defect, (a) the captured image itself, (b) a high-resolution image obtained by the 2 × 2 pixel shift method, (c) a high resolution obtained by the 1 × 2 pixel shift method The reproducibility (stability) of defect detection based on each of the image and (d) the high-resolution image obtained by the 2 × 1 pixel shift method is compared. As a measure of reproducibility, a variation in defect detection size caused by a minute movement of the reference imaging position is used.

  The image obtained by imaging the evaluation object 200 from the position shown in the upper left part of FIG. 9A is binarized (1 is assigned to a pixel in which the area occupied by the defect is more than half, and 0 is assigned to the other pixel. (Corresponding to assigning), the binarized image shown in the lower left of FIG. 9A is obtained. When the defect size is counted based on the binarized image, the detected size is 3 pixels. On the other hand, when the image obtained by imaging the evaluation object 200 from the position shown in the upper right part of FIG. 9A is binarized, the binarized image shown in the lower right part of FIG. 9A is obtained. When the defect size is counted based on the binarized image, the detected size is 2 pixels. As described above, when a defect is detected using the captured image itself, the detection size changes greatly due to the movement of the imaging position, and the reproducibility is low.

  FIG. 9B shows a binarized image obtained by binarizing a high-resolution image obtained by the 2 × 2 pixel shift method. When a high-resolution image obtained by the 2 × 2 pixel shifting method is used, the effective pixel size is a quarter (area ratio) of the original pixel size. For this reason, the defect size counted based on the binarized image is unchanged (all 9 effective pixels) before and after the movement. That is, the reproducibility is higher than when a defect is detected using the captured image itself.

  However, in order to generate a high-resolution image by the 2 × 2 pixel shifting method, four images are required, and the time required for imaging and the time required for the high-resolution processing become problems. Therefore, use of the 1 × 2 pixel shift method or the 2 × 1 pixel shift method is effective.

  FIG. 9C shows a binarized image obtained by binarizing a high-resolution image obtained by the 1 × 2 pixel shift method. FIG. 9D shows a binarized image obtained by binarizing a high resolution image obtained by the 2 × 1 pixel shift method.

  When a high resolution image obtained by the 1 × 2 pixel shift method is used to detect a vertically long defect, as shown in FIG. 9C, the defect size counted based on the binarized image is , Change before and after the movement (decrease by one execution pixel). On the other hand, when a high-resolution image obtained by the 2 × 1 pixel shift method is used, the defect size counted based on the binarized image does not change before and after the movement (see FIG. 9D). All are 4 effective pixels). That is, when detecting a vertically long defect, the 2 × 1 pixel shift method for improving the horizontal resolution is effective in improving the reproducibility of defect detection. Similarly, when detecting a horizontally long defect, a 1 × 2 pixel shift method for improving the vertical resolution is effective in improving the reproducibility of defect detection.

  FIG. 10 shows (a) the captured image itself, (b) a high-resolution image obtained by the 2 × 2 pixel shift method, (c) a high-resolution image obtained by the 1 × 2 pixel shift method, and ( d) A graph showing the variation in defect intensity calculated based on each of the high-resolution images obtained by the 2 × 1 pixel shifting method. Each plot point is obtained by capturing an image necessary for each case on a liquid crystal display panel having a vertically long defect while slightly changing the reference imaging position. The contrast volume described above was used for the evaluation of the defect strength.

  In the case of using the picked-up image itself and a high-resolution image obtained by the 1 × 2 pixel shifting method, the defect intensity varies greatly, whereas 2 × 1 or 2 × 2 pixels When a high-resolution image obtained by the shifting method is used, the variation in defect strength is small. This result agrees with the result shown in FIG. 9 that it is effective to improve the horizontal resolution by using the 2 × 1 pixel shifting method for the vertically long defect.

  As described above, if the high resolution processing is performed using two images captured by shifting the imaging position in the minor axis direction of the defect, it is obtained when the high resolution processing is performed using four images. Reproducibility that is almost the same as reproducibility can be obtained, and the number of images for high resolution processing can be halved.

(Program and recording medium)
Finally, each block included in the PC 120 may be configured by hardware logic, or may be realized by software using a CPU as follows.

  That is, the PC 120 includes a CPU (central processing unit) that executes instructions of a defect detection program that realizes each function, a ROM (read only memory) that stores the program, a RAM (random access memory) that develops the program, A storage device (recording medium) such as a memory for storing programs and various data is provided. An object of the present invention is to supply a PC 120 with a recording medium in which a program code (execution format program, intermediate code program, source program) of a defect detection program, which is software for realizing the functions described above, is recorded so as to be readable by a computer. This can also be achieved by the PC 120 reading and executing the program code recorded on the recording medium.

  Examples of the recording medium include a tape system such as a magnetic tape and a cassette tape, a magnetic disk such as a floppy (registered trademark) disk / hard disk, and an optical disk such as a CD-ROM / MO / MD / DVD / CD-R. Card system such as IC card, IC card (including memory card) / optical card, or semiconductor memory system such as mask ROM / EPROM / EEPROM / flash ROM.

  Further, the PC 120 may be configured to be connectable to a communication network, and the program code may be supplied to the PC 120 via the communication network. The communication network is not particularly limited. For example, the Internet, intranet, extranet, LAN, ISDN, VAN, CATV communication network, virtual private network, telephone line network, mobile communication network, satellite communication. A net or the like is available. Further, the transmission medium constituting the communication network is not particularly limited. For example, even in the case of wired such as IEEE 1394, USB, power line carrier, cable TV line, telephone line, ADSL line, etc., infrared rays such as IrDA and remote control, Bluetooth ( (Registered trademark), 802.11 wireless, HDR, mobile phone network, satellite line, terrestrial digital network, and the like can also be used. The present invention can also be realized in the form of a computer data signal embedded in a carrier wave in which the program code is embodied by electronic transmission.

(Additional notes)
The present invention is not limited to the above-described embodiments, and various modifications can be made within the scope of the claims. That is, embodiments obtained by combining technical means appropriately modified within the scope of the claims are also included in the technical scope of the present invention.

  For example, in the above-described embodiment, the pixel shifting method is used as the resolution enhancement processing. However, the present invention is not limited to this, and high resolution processing other than the pixel shifting method, for example, super-resolution processing is performed. It may be used.

  As an example of super-resolution processing, the main points of reconstruction-type super-resolution processing using ML (Maximum-likelihood) method, MAP (Maximum A Posterior) method, POCS (Projection On to Convex Sets) method, etc. Is described as follows.

  That is, in the reconstruction type super-resolution processing, a plurality of low resolution images corresponding to the imaging model are estimated from the assumed high resolution image. For estimation of the low resolution image, for example, a point spread function (PSF function) determined by the imaging model is used. Then, the high-resolution image is reasserted so that the difference between the estimated low-resolution image and the actual captured image becomes small. This is repeated until the difference between the estimated low resolution image and the actual captured image converges to obtain a high resolution image. It is clear that the object of the present invention can be achieved even if the number of captured images referred to in such super-resolution processing is set based on defect information indicating the shape of the defect to be detected.

  For example, the present invention can also be expressed as follows.

  1. an imaging unit that images an evaluation target, an imaging timing control unit that controls an imaging timing of the imaging unit, a motion control unit that controls a motion of the imaging unit, and a plurality of images captured by the imaging unit High resolution processing means for generating a high resolution image from an image, a defect shape information holding unit for holding defect shape information of a target defect, and the number of images used for high resolution processing according to the defect shape information holding unit A defect inspection apparatus comprising: a high-resolution processing selection unit that fluctuates, and a defect detection unit that inspects a high-resolution defect image.

  2. 2. The defect inspection apparatus according to 1, wherein a pixel shifting method is used as the high resolution processing means.

  3. The defect inspection apparatus according to 1, wherein the defect shape information holding unit obtains and inputs defect appearance tendency information to be evaluated in advance.

  4). 2. The defect inspection apparatus according to 1, wherein in the defect shape information holding unit, defect appearance tendency information to be evaluated is sequentially updated according to an inspection result.

  5). 2. The defect inspection apparatus according to 1, wherein the high-resolution processing selection unit varies the image position / number of images used for the high-resolution processing in accordance with the defect aspect ratio.

  6). In the high-resolution processing selection means, a case where there is no 1 × 1 high-resolution processing and a case where a 1 × 2, 2 × 1, 2 × 2 pixel shifting method is used based on the defect aspect ratio. 2. The defect inspection apparatus according to 1, wherein selection is performed and high resolution is performed.

  7. 2. The defect detection unit according to 1, wherein a contrast is calculated from a difference from a luminance value of a peripheral pixel with respect to a target pixel of a defect image, and a defect is detected when the contrast is equal to or greater than a threshold value. Defect inspection equipment.

  8). In the defect detection means, when the evaluation target has a periodic pattern, the contrast is calculated from the difference between the target pixel pattern and a neighboring pixel having the same pattern, and is detected as a defect when the contrast is equal to or greater than a certain threshold value. 2. The defect inspection apparatus according to 1, wherein

  9. 2. The defect inspection apparatus according to 1, wherein the defect detection means calculates a contrast volume obtained by integrating the contrast of the defect portions and evaluates the strength of the defect.

  The present invention can be used to detect various defects in various evaluation objects. The type of defect may be anything that can be captured as an image in the captured image, and the type of the evaluation object may be anything as long as it can contain such a defect. In particular, it can be suitably used for detecting defects such as black spot defects and bright spot defects in display devices such as liquid crystal display panels.

1, showing an embodiment of the present invention, is a block diagram illustrating a configuration of a defect detection apparatus. FIG. FIG. 3 is a flowchart illustrating an operation example of the defect detection apparatus according to the embodiment of the present invention. It is a flowchart which shows embodiment of this invention and shows the other operation example of a defect detection apparatus. 4 is a flowchart illustrating an imaging position setting process according to an exemplary embodiment of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 illustrates an embodiment of the present invention and is an explanatory diagram illustrating an outline of a resolution enhancement process using a 2 × 1 pixel shift method. FIG. 9 is an explanatory diagram illustrating an embodiment of the present invention and an overview of a resolution enhancement process using a 2 × 2 pixel shift method. FIG. 8 is a diagram illustrating an embodiment of the present invention and is a diagram illustrating a contrast calculation method that is preferable when a defect size corresponds to one pixel of a high-resolution image. FIG. 9 is a diagram illustrating an embodiment of the present invention and explaining a contrast calculation method suitable for a case where a defect size corresponds to several pixels of a high-resolution image. FIG. 4 is a diagram illustrating an embodiment of the present invention and illustrating the effectiveness of a 2 × 1 pixel shifting method and a 1 × 2 pixel shifting method. 4 is a graph showing an embodiment of the present invention and comparing reproducibility of defect detection accuracy by each pixel shifting method.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Defect detection apparatus 110 Optical system 111 Camera 112 Camera drive apparatus 113 Stage 114 Stage drive apparatus 120 PC
121 recording unit 122 optical system control unit 123 high resolution processing unit 124 defect detection processing unit

Claims (12)

Generating means for referring to a plurality of captured images obtained by imaging the evaluation object and generating a high-resolution image having a resolution higher than the resolution of the captured image;
Detecting means for detecting a defect in the evaluation object with reference to the high resolution image generated by the generating means;
Setting means for setting the number of captured images referred to by the generating means to generate the resolution-enhanced image based on defect shape information indicating the shape of the defect to be detected;
A defect detection apparatus comprising: The generating means generates the high-resolution image by a pixel shifting method.
The defect detection apparatus according to claim 1. It further comprises a recording unit that records defect shape information indicating the shape of the defect estimated in advance,
The setting unit is configured to set the number of captured images referred to by the generation unit to generate the high-resolution image based on the defect shape information recorded in the recording unit.
The defect detection apparatus according to claim 1, wherein the defect detection apparatus is a defect detection apparatus. It further comprises a recording unit for recording defect shape information indicating the shape of the defect detected by the defect detection means,
The setting unit is configured to set the number of captured images referred to by the generation unit to generate the high-resolution image based on the defect shape information recorded in the recording unit.
The defect detection apparatus according to claim 1, wherein the defect detection apparatus is a defect detection apparatus. The setting means sets, based on the aspect ratio of the defect to be detected, whether to generate a resolution-enhanced image with reference to a captured image obtained by capturing the evaluation object in which direction. To do,
The defect detection apparatus according to claim 2. The setting means shifts two evaluation images obtained by shifting the evaluation object in the first direction and shifting the evaluation object in a second direction perpendicular to the first direction. Either of the two captured images obtained by imaging or the four captured images obtained by shifting the evaluation object in the first direction and / or the second direction. To set whether to generate the high resolution image with reference to
The defect detection apparatus according to claim 5. The detection means calculates a contrast of each pixel constituting the high-resolution image, and identifies a set of pixels in which the calculated contrast exceeds a predetermined threshold as a defect image.
The defect detection apparatus according to claim 1, wherein the defect detection apparatus is a defect detection apparatus. The detection means is a reference pixel determined according to the pixel value of the pixel and the pattern of the evaluation object, with the contrast of each pixel constituting the high resolution image, and if there is no defect, the pixel Is calculated by taking the absolute value of the difference from the average pixel value of the reference pixels estimated to have the same pixel value as
The defect detection apparatus according to claim 7. The detection means calculates the strength of the defect by integrating the contrast of each pixel constituting the defect image.
The defect detection apparatus according to claim 7 or 8, wherein A generation step of referring to a plurality of captured images obtained by imaging the evaluation object by the imaging device and generating a high-resolution image having a higher resolution than the resolution of the imaging device;
With reference to the high resolution image generated in the generation step, a detection step of detecting a defect in the evaluation object,
A setting step for setting the number of the captured images referred to in the generation step in order to generate the resolution-enhanced image based on the defect shape information indicating the shape of the defect to be detected. A defect detection method characterized by the above.   A defect detection program for causing a computer to operate as the defect detection apparatus according to claim 1, wherein the computer functions as each of the units included in the defect detection apparatus.   A computer-readable recording medium on which the defect detection program according to claim 11 is recorded.

Images