JP6688629B2 - Defect detecting device, defect detecting method and program - Google Patents

Description

  The present invention relates to a technique for detecting defects on the surface of an object.

  2. Description of the Related Art Conventionally, there has been used a device that irradiates an object with light to capture an image and inspects the appearance of the object based on the captured image. In such an inspection apparatus, when detecting a defect that is a concave portion or a convex portion, it is necessary to prevent the surface color due to dirt or the like from being detected as a defect.

  For example, in the defect inspection apparatus for a sheet-shaped material of Patent Document 1, the sheet-shaped material is irradiated with light from an oblique direction, and the concave portion, the convex portion, and the dirty portion are classified based on the positional relationship between the dark shadow area and the bright bright spot area. To be done. The appearance inspection apparatus of Patent Document 2 changes the illuminance of the surface to be inspected, and utilizes the fact that the change in density in the image due to the change in illuminance is different between the flaw and the oil, thereby accurately detecting the flaw. In the rod-shaped object inspection device of Patent Document 3, a rod-shaped object is irradiated with red light and blue light from opposite directions, a color image is decomposed into a red image and a blue image, and positions of bright and dark regions of each image are separated. The convex defect and the stain are detected from the relationship.

JP-A-2-156144 JP, 2002-116153, A JP, 2005-37203, A

  By the way, due to the unevenness of the surface of the object, a dark area or a bright area appears in the image in a complicated manner, and it may not be possible to accurately separate the unevenness defect and the stain as in the prior art.

  The present invention has been made in view of the above problems, and it is an object of the present invention to suppress overdetection of defects by removing with high accuracy a defect caused by the surface color of an object.

The invention according to claim 1 is a defect detection device for detecting a defect on a surface of an object, and a light irradiation unit capable of irradiating the object with light in a plurality of differently biased illumination states. an imaging unit that acquires an image of the target area of the surface of the object as a captured image, an illumination state by the light irradiation unit, an imaging control unit that controls the acquisition of images by the imaging unit, the plurality of illumination From at least one captured image acquired in at least one illumination state included in the state and used for defect detection, refer to a corresponding reference image and defect one of a dark region and a bright region with respect to the reference image. obtained as candidate regions, from each of the plurality of captured images acquired by two or more lighting states included in the plurality of lighting conditions, darker with respect to the reference image while referring to the reference image corresponding The other of the area and the bright area is obtained as a light and dark reverse area, and among the defect candidate areas, one that does not overlap with any light and dark reverse area by a predetermined condition or more is excluded from the defect candidates, and then a defect is generated based on the defect candidate area. And a defect acquisition unit for acquiring the presence of the defect.

The invention according to claim 2 is the defect detecting device according to claim 1, wherein the defect detecting section uses each of a plurality of captured images acquired in the plurality of illumination states for defect detection. sequentially selected as being that IMAGING image is.

The invention according to claim 3 is the defect detecting apparatus according to claim 1 or 2, wherein the at least one captured image used for defect detection includes a first captured image and a second captured image. The defect acquisition unit acquires, from the first captured image, one of a dark region and a bright region with respect to the first reference image as the defect candidate region while referring to the corresponding first reference image, 2 captured image, obtains the other dark area and bright area relative to the second reference picture with reference to the corresponding second reference image as another defect candidate regions, light dark that corresponds to the defect candidate regions when acquiring the reverse area, wherein the other defect candidate area from the second image is acquired, when obtaining the brightness inverse area corresponding to the other defect candidate regions, from the first captured image The defect candidate area is acquired.

  A fourth aspect of the present invention is the defect detection apparatus according to any one of the first to third aspects, wherein the defect acquisition section uses a first image from the at least one captured image utilized for defect detection. A first defect candidate area is acquired by a method, a second defect candidate area is acquired by a second method different from the first method, and based on the first defect candidate area and the second defect candidate area. The defect candidate area is acquired.

  A fifth aspect of the present invention is the defect detecting apparatus according to the fourth aspect, wherein the first method performs alignment between the captured image and the reference image, and then detects the difference image from these images. The second method is a method of acquiring the first defect candidate area, and the second method performs the minute area removal processing on the captured image and the reference image, and then extracts the second defect candidate area from the difference image of these images. This is the method of acquisition.

The invention according to claim 6 is a defect detection method for detecting defects on the surface of an object, wherein a) while the object is being irradiated with light in each of a plurality of differently biased illumination states. A step of acquiring a plurality of picked-up images by acquiring an image of a target area on the surface of the target object by an imaging unit, and b) being acquired in at least one illumination state included in the plurality of illumination states. Acquiring one of a dark region and a bright region with respect to the reference image as a defect candidate region from at least one captured image used for defect detection while referring to a corresponding reference image; and c) the plurality of illuminations from each of two or more captured images acquired by the lighting conditions that are included in the state, the other a dark reverse dark areas and light areas with respect to the reference image while referring to the reference image corresponding A step of acquiring as a region, and d) out of the defect candidate regions, which do not overlap any of the light-dark reverse regions by a predetermined condition or more, is excluded from the defect candidates, and the existence of the defect is acquired based on the defect candidate regions. And a process.

The invention described in claim 7 is the defect detection method according to claim 6, each of the previous SL plural been multiple captured image acquired in the illuminating state, captured images used for defect detection The steps b) to d) are repeated while sequentially selecting as .

The invention according to claim 8 is the defect detection method according to claim 6 or 7, wherein the at least one captured image used for defect detection includes a first captured image and a second captured image. In the step b), one of a dark region and a bright region with respect to the first reference image is acquired as the defect candidate region from the first captured image while referring to the corresponding first reference image, and the defect is detected. detection method, e) from said second captured image, to get the corresponding other of the dark areas and light areas with respect to the second reference picture with reference to the second reference image as another defect candidate region step further When the bright / dark reverse region corresponding to the defect candidate region is acquired, the other defect candidate region is acquired from the second captured image, and the bright / dark reverse region corresponding to the other defect candidate region is acquired. In front The defect candidate area is acquired from the first captured image.

The invention according to claim 9 is a program for causing a computer to detect a defect in the target region from a plurality of images of the target region on the surface of the target, wherein the execution of the program by the computer causes the computer to: a) preparing a plurality of captured images of the target area and corresponding reference images acquired in a plurality of differently biased illumination states, and b) at least one illumination state included in the plurality of illumination states. And acquiring one of a dark region and a bright region with respect to the reference image as a defect candidate region from at least one captured image acquired in step S1 and used for defect detection, c ) from each of the plurality of captured images acquired by two or more lighting states included in the plurality of lighting conditions, corresponding reference A step of acquiring the other of the dark area and the bright area as a bright / dark reverse area with respect to the reference image while referring to the image; and d) selecting one of the defect candidate areas that does not overlap any of the bright / dark reverse areas by a predetermined condition or more. After removing from the defect candidates, the step of acquiring the presence of the defect based on the defect candidate area is executed.

  According to the present invention, overdetection of defects can be suppressed.

It is a figure which shows the structure of a defect detection apparatus. It is a top view which shows the main body of a defect detection apparatus. It is a figure which shows the structure of a computer. It is a block diagram which shows the function structure implement | achieved by a computer. It is a figure which shows the structure of a 1st dark defect candidate acquisition part. It is a figure which shows the structure of a 2nd dark defect candidate acquisition part. It is a figure which shows the structure of a 2nd bright defect candidate acquisition part. It is a figure which shows the structure which acquires dark defect candidate data. It is a figure which shows the structure which limits dark defect candidate data. It is a figure which shows the flow of operation | movement of a defect detection apparatus. It is a figure which illustrates a captured image. It is a figure which illustrates a reference image. It is a figure which shows a 1st dark defect candidate image. It is a figure which shows the captured image after performing expansion and contraction processing. It is a figure which shows the reference image after performing expansion and contraction processing. It is a figure which shows the 2nd dark defect candidate image. It is a figure which shows a dark defect candidate image. It is a figure which illustrates a bright / dark reverse region image. It is a figure which illustrates another bright / dark reverse region image. It is a figure which shows a logical sum image. It is a figure which overlaps a dark defect candidate image and a logical sum image.

  FIG. 1 is a diagram showing a configuration of a defect detection device 1 according to an embodiment of the present invention. FIG. 2 is a plan view showing the main body 11 of the defect detection device 1. The defect detection device 1 is a device that inspects the appearance of a three-dimensional object 9 having a surface that is not a mirror surface, and detects defects on the surface of the object 9. The target 9 is, for example, a metal component formed by forging or casting. The surface of the object 9 is in a satin-like shape having fine irregularities, that is, in a matte state. The surface of the object 9 is processed by shot blasting such as sand blasting. The object 9 is, for example, various parts used for a universal joint. The surface of the object 9 may be glossy as long as it scatters light to some extent.

  The defect on the surface of the object 9 is a portion that is concave or convex with respect to the ideal shape. The defect is, for example, a dent, a scratch, a processing defect, or the like.

  As shown in FIG. 1, the defect detection device 1 includes a main body 11 and a computer 12. The main body 11 includes a holding unit 2, a plurality of image pickup units 3 (in FIG. 1, reference numerals 3a, 3b, and 3c are attached, but if they are not distinguished, a reference numeral 3 is attached), and a light irradiation unit 4. . The object 9 is held by the holding unit 2. The main body 11 is provided with a light-shielding cover (not shown) that prevents external light from reaching the holding unit 2, and the holding unit 2, the imaging unit 3, and the light irradiation unit 4 are provided inside the light-shielding cover.

  When automatically inspecting the entire surface of the object 9, another body 11 is provided. A mechanism is provided between the two main bodies 11 to turn the object 9 upside down and convey the object 9.

  As shown in FIGS. 1 and 2, the plurality of image capturing units 3 include one upper image capturing unit 3a, eight oblique image capturing units 3b, and eight side image capturing units 3c. The upper image pickup unit 3 a is arranged above the holding unit 2. An image of the object 9 on the holding unit 2 taken from directly above can be acquired by the upper imaging unit 3a.

  As shown in FIG. 2, when the main body 11 is viewed from the upper side downward (that is, when the main body 11 is viewed in plan), the eight oblique imaging units 3 b are arranged around the holding unit 2. It The eight oblique imaging units 3b pass through the center of the holding unit 2 and are arranged at an angular interval (angle pitch) of 45 ° in the circumferential direction about the central axis J1 that faces the up-down direction. As shown in FIG. 1, the angle θ2 formed by the optical axis K2 and the central axis J1 is about 45 ° in the plane including the optical axis K2 and the central axis J1 of each oblique imaging unit 3b. An image of the object 9 on the holding unit 2 taken obliquely from above can be acquired by each oblique imaging unit 3b. The angle θ2 is not limited to 45 ° as long as the object 9 is imaged obliquely from above, and may be arbitrarily set within the range of 15 to 75 °.

  When the main body 11 is viewed in a plan view, the eight side image capturing units 3c are also arranged around the holding unit 2 like the eight oblique image capturing units 3b. The eight side imaging units 3c are arranged in the circumferential direction at angular intervals of 45 °. An angle θ3 formed by the optical axis K3 and the central axis J1 is about 90 ° in a plane including the optical axis K3 and the central axis J1 of each side imaging unit 3c. An image of the object 9 on the holding unit 2 taken from the side can be acquired by each side imaging unit 3c.

  The upper image pickup unit 3a, the oblique image pickup unit 3b, and the side image pickup unit 3c have a two-dimensional image sensor such as a CCD (Charge Coupled Device) or a CMOS (Complementary Metal-Oxide Semiconductor), and have a multi-gradation image. Is obtained. The upper imaging unit 3a, the oblique imaging unit 3b, and the lateral imaging unit 3c are supported by a supporting unit (not shown).

  The light irradiation section 4 includes one upper light source section 4a, eight oblique light source sections 4b, and eight side light source sections 4c. The upper light source unit 4a is adjacent to the upper imaging unit 3a. In the upper light source unit 4a, a plurality of LEDs (light emitting diodes) are arranged vertically to the central axis J1, that is, horizontally. The upper light source unit 4a irradiates the object 9 on the holding unit 2 with light from almost directly above.

  When the main body 11 is viewed in a plan view, the eight oblique light source units 4 b are arranged around the holding unit 2. The oblique light source units 4b are adjacent to the oblique imaging unit 3b. The eight oblique light source units 4b are arranged at an angular interval of 45 ° in the circumferential direction. In each oblique light source section 4b, a plurality of LEDs are arranged substantially perpendicular to the optical axis K2. In each of the oblique light source parts 4b, the object 9 on the holding part 2 can be irradiated with light obliquely from above.

  When the main body 11 is viewed in a plan view, the eight side light source units 4 c are arranged around the holding unit 2. The lateral light source units 4c are adjacent to the lateral imaging unit 3c, respectively. The eight side light source portions 4c are arranged at an angular interval of 45 ° in the circumferential direction. In each side light source unit 4c, a plurality of LEDs are arranged substantially perpendicular to the optical axis K3 and horizontally. Therefore, the eight lateral light source portions 4c have a substantially octagonal shape in a plan view. In each side light source unit 4c, the object 9 on the holding unit 2 can be laterally irradiated with light. Light sources other than LEDs may be used in the upper light source unit 4a, the oblique light source unit 4b, and the side light source unit 4c. Although the color of light is white in the present embodiment, the color of light and the wavelength band may be variously changed. The light irradiation unit 4 can irradiate the object 9 with diffused light from various directions. Hereinafter, each state in which light is emitted from at least one specific light source unit and the object 9 is illuminated with the polarized light is referred to as an “illumination state”. The light irradiator 4 can irradiate the object 9 with light in a plurality of differently biased illumination states.

  FIG. 3 is a diagram showing the configuration of the computer 12. The computer 12 has a general computer system configuration including a CPU 121 that performs various arithmetic processes, a ROM 122 that stores a basic program, and a RAM 123 that stores various types of information. The computer 12 has a fixed disk 124 for storing information, a display 125 for displaying various kinds of information such as images, a keyboard 126a and a mouse 126b for accepting inputs from an operator (hereinafter collectively referred to as "input unit 126"), A reader 127 that reads information from a computer-readable recording medium 8 such as an optical disk, a magnetic disk, and a magneto-optical disk, and a communication unit 128 that transmits / receives signals to / from other components of the defect detection apparatus 1 are further provided. Including.

  In the computer 12, the program 80 is previously read from the recording medium 8 via the reading device 127 and stored in the fixed disk 124. The CPU 121 executes arithmetic processing according to the program 80 while using the RAM 123 and the fixed disk 124. The CPU 121 functions as a calculation unit in the computer 12. Other than the CPU 121, another configuration that functions as a calculation unit may be adopted.

  FIG. 4 is a diagram showing functions realized by the computer 12 executing arithmetic processing according to the program 80. In FIG. 4, the imaging control unit 51, the defect acquisition unit 52, and the storage unit 53 correspond to the functions realized by the computer 12. All or part of these functions may be realized by a dedicated electric circuit. Further, these functions may be realized by a plurality of computers.

  The imaging control unit 51 controls the imaging unit 3 and the light irradiation unit 4, and acquires an image of the object 9 (to be exact, data indicating the image). The image data is stored in the storage unit 53. Although the image pickup unit 3 is shown as one block in FIG. 4, the upper image pickup unit 3a, the oblique image pickup unit 3b, and the side image pickup unit 3c are actually connected to the image pickup control unit 51.

  As will be described later, every time the imaging control unit 51 controls each light source unit of the light irradiation unit 4 to change the illumination state, image data is acquired by at least one of the 17 imaging units 3. Hereinafter, the image acquired by imaging will be referred to as “captured image”, and the data will be referred to as “captured image data”. The captured image data 911 is stored in the storage unit 53. Data of an image of the ideal object 9 in each illumination state is stored in the storage unit 53 as reference image data 912. That is, ideal image data corresponding to each illumination state of each imaging unit 3 is prepared in the storage unit 53 as reference image data 912.

  The illumination state of light by the light irradiation unit 4 refers to a state in which the object 9 is irradiated with light from a specific irradiation direction. The irradiation direction is not strictly defined, and means the approximate irradiation direction of light. The irradiation direction is not limited to the parallel light when the light is emitted only from that direction. The fact that the light is emitted from a certain direction means that the light is emitted with a deviation from that direction. Further, the number of light irradiation directions in one shooting is not limited to one. For example, light may be simultaneously emitted from a plurality of light source units that are separated from each other.

  The defect acquisition unit 52 includes a first dark defect candidate acquisition unit 521, a second dark defect candidate acquisition unit 522, a first bright defect candidate acquisition unit 523, and a second bright defect candidate acquisition unit 524. The “dark defect” means a defect that appears dark in an image. "Bright defect" means a defect that appears bright in an image. FIG. 5 is a diagram showing the configuration of the first dark defect candidate acquisition unit 521. The first dark defect candidate acquisition unit 521 includes two filter processing units 531, a pre-alignment unit 532, a shake comparison unit 533, a binarization unit 534, and an area filter unit 535. The first dark defect candidate acquisition unit 521 acquires the first dark defect candidate data 921 which is image data indicating the first dark defect candidate area. The first dark defect candidate area is an area dark in the captured image with respect to the reference image.

  FIG. 6 is a diagram showing the configuration of the second dark defect candidate acquisition unit 522. In FIG. 6, the upstream side from the pre-alignment unit 532 is common to the first dark defect candidate acquisition unit 521. The second dark defect candidate acquisition unit 522 includes an expansion processing unit 541, a contraction processing unit 542, a comparison unit 543, a binarization unit 544, and an area filter unit 545, following the pre-alignment unit 532. The second dark defect candidate acquisition unit 522 acquires second dark defect candidate data 922 that is image data indicating the second dark defect candidate area. The second dark defect candidate area is an area dark in the captured image with respect to the reference image. As described later, the method of acquiring the first dark defect candidate area is different from the method of acquiring the second dark defect candidate area.

  FIG. 7 is a diagram showing the configuration of the second bright defect candidate acquisition unit 524. In FIG. 7, the upstream side from the pre-alignment unit 532 is common to the first dark defect candidate acquisition unit 521. The second bright defect candidate acquisition unit 524 includes a contraction processing unit 551, an expansion processing unit 552, a comparison unit 553, a binarization unit 554, and an area filter unit 555, following the pre-alignment unit 532. The second bright defect candidate acquisition unit 524 acquires bright / dark reverse region data 923, which is image data indicating a bright / dark reverse region described later.

  When performing the processing at high speed, a large number of the first dark defect candidate acquisition unit 521, the second dark defect candidate acquisition unit 522, the first bright defect candidate acquisition unit 523, and the second bright defect candidate acquisition unit 524 are used as the defect acquisition unit 52. Is provided at the same time, and processing is performed in parallel on a plurality of captured images.

  FIG. 8 is a diagram showing a configuration in which the defect acquisition unit 52 acquires dark defect candidate data 924, which is image data indicating a dark defect candidate region, from the first dark defect candidate data 921 and the second dark defect candidate data 922. is there. This configuration includes a plurality of AND operation units 561 and a region selection unit 562. FIG. 9 is a diagram showing a configuration in which the defect acquisition unit 52 limits the dark defect candidate area indicated by the dark defect candidate data 924 by using the bright / dark reverse area data 923. This configuration includes a logical sum operation unit 563 and a candidate limiting unit 564.

  FIG. 10 is a diagram showing a flow of operations of the defect detection apparatus 1. First, the object 9 to be inspected is held on the holder 2. The holding portion 2 is provided with, for example, an abutting portion for alignment, and when a predetermined portion of the object 9 comes into contact with the abutting portion, the object 9 is oriented at a predetermined position in a predetermined direction. Will be placed in. The holding unit 2 may be a stage provided with a positioning pin.

  Next, the image pickup control unit 51 changes the illumination state by changing the light source unit to be turned on, and the image pickup is performed by the selected image pickup unit 3 (steps S11 to S13). Specifically, one side image pickup unit 3c is selected, and one of the five side light source units 4c that are horizontally continuous with the side image pickup unit 3c as the center is sequentially selected and turned on. Then, the side imaging unit 3c acquires an image. The above operation is repeated while changing the side imaging unit 3c. Actually, the operation time can be shortened by capturing images by the plurality of side image capturing units 3c in each illumination state. Further, all the side light source units 4c are turned on and images are taken by all the side image pickup units 3c. As a result, six images are acquired by each side imaging unit 3c.

  In the case of the oblique image capturing unit 3b, one oblique image capturing unit 3b is selected, one of the eight oblique light source units 4b is sequentially selected and turned on, and the oblique image capturing unit 3b acquires an image each time it is turned on. To do. The above operation is repeated while changing the oblique imaging unit 3b. Actually, the operation time can be shortened by performing image capturing by all the oblique image capturing units 3b in each illumination state. Further, all the oblique light source units 4b are turned on, and images are taken by all the oblique imaging units 3b. Imaging is performed by all the oblique imaging units 3b even when only the upper light source unit 4a is turned on. As a result, 10 images are acquired by each oblique imaging unit 3b.

  In the case of the upper imaging unit 3a, the illumination state is changed similarly to the oblique imaging unit 3b and 10 images are acquired. In an actual operation, the operation time can be shortened by performing image pickup by the upper image pickup unit 3a at the time of image pickup by the oblique image pickup unit 3b.

  The data of the captured image is stored in the storage unit 53 as captured image data 911. As described above, the reference image data corresponding to each captured image is prepared as the reference image data 912 in the storage unit 53. The reference image shows the object 9 having no defect under the same illumination condition as the captured image. The reference image data 912 may be acquired by capturing an image of the target object 9 having no defect, or may be acquired as data of an average image of images of a large number of target objects 9.

  Hereinafter, in order to simplify the expression, the processing on the image data may be simply expressed as the processing on the image. Further, only the processing focusing on one image pickup unit 3 will be described. Similar processing is performed on the other image pickup units 3. The defect detection apparatus 1 can detect the presence of a dark defect and the presence of a bright defect, but the following description focuses on the dark defect. Further, the “dark defect candidate area” refers to an area that appears dark as a candidate indicating the presence of a defect in the image.

  First, one captured image is selected, and the reference image corresponding to the captured image is selected. As shown in FIG. 5, in the first dark defect candidate acquisition unit 521, the captured image data 911 of the captured image and the reference image data 912 of the reference image are input to the filter processing unit 531.

  In the two filter processing units 531, filter processing for reducing noise such as a median filter and a Gaussian filter is performed on the captured image and the reference image, respectively. The filtered captured image and the reference image are output to the pre-alignment unit 532. The pre-alignment unit 532 identifies the relative position and angular deviation of the reference image with respect to the captured image by pattern matching using a predetermined pattern. Then, the reference image is translated and rotated with respect to the captured image by an amount of positional and angular displacement between the two images, so that the approximate position and angle of the reference image are matched with the captured image. As a result, pre-alignment is performed on both images.

  The shake comparison section 533 obtains an evaluation value indicating the difference between the captured image and the reference image while moving the reference image from the pre-aligned position little by little in the vertical and horizontal directions. As the evaluation value, for example, the sum of the absolute values of the differences (with signs) of the pixel values in the area where both images overlap is obtained. Then, an image showing the signed difference in the pixel values of both images at the position where the evaluation value is the minimum is acquired. The signed difference image is binarized with a predetermined value, and a first dark defect candidate image indicating the first dark defect candidate area is acquired.

  In reality, the signed difference image is not obtained in order to simplify the processing. Specifically, the value of the corresponding pixel of the captured image is subtracted from the value of each pixel of the reference image, and the value of the pixel of the difference image is obtained by setting the value to 0 when the value is negative. A positive value is prepared in advance, and an area formed by pixels having a positive value or more in the difference image is acquired as the first dark defect candidate area. Generally speaking, a region having a lower brightness in the first captured image than the first reference image and an absolute value of the difference in brightness not less than the first reference value is acquired as the first dark defect candidate region. . The first reference value is a positive value. In other words, a region of the captured image whose brightness is lower than the reference image by the predetermined value or more is acquired as the first dark defect candidate region. When the image is monochrome, the pixel value may be regarded as the lightness, and in the case of a color image, the value obtained by performing a predetermined calculation on the pixel value of each color component is treated as the lightness.

  The first dark defect candidate area may be obtained from the ratio of the value of each pixel of the reference image and the value of the corresponding pixel of the captured image. Specifically, the value of each pixel of the reference image is divided by the value of the corresponding pixel of the captured image to obtain the pixel value of the ratio image. A first reference value larger than 1 is prepared in advance, and an area formed by pixels having a value equal to or larger than the first reference value in the ratio image is acquired as the first dark defect candidate area. Of course, the value of the pixel of the ratio image may be obtained by dividing the value of each pixel of the captured image by the value of the corresponding pixel of the reference image. In this case, a region formed by pixels having a value equal to or smaller than the first reference value smaller than 1 in the ratio image is acquired as the first dark defect candidate region.

  The first reference value does not have to be a constant. The first reference value may be a function of the brightness or pixel value of the reference image and / or the captured image. The first reference value may be determined using the difference or ratio between the brightness or the pixel value of the reference image and the captured image, and still another calculation may be used. The point that the first reference value may be variously set is the same for the second and third reference values described later. The first to third reference values do not have to be the same value, and the calculation method may be different. Generally speaking, in the captured image, a region whose brightness is lower than that of the reference image and lower than a value satisfying a predetermined condition is acquired as the first dark defect candidate region. The “predetermined condition” may be set individually for each captured image. Furthermore, a plurality of “predetermined conditions” may be used for one captured image. For example, the first reference value may be set so that it is hard to be detected as a defect candidate region at a position where the pixel value is likely to change from image to image, such as an edge in a captured image. The above description also applies to the following second dark defect candidate region, first bright defect candidate region region, and second bright defect candidate region.

  FIG. 11A is a diagram illustrating a captured image 811. FIG. 11B is a diagram illustrating the reference image 812. Hereinafter, a region of the surface of the target object 9 that appears in the captured image will be referred to as a “target region 70”. The image capturing units 3 and the target regions 70 correspond one-to-one, and each image capturing unit 3 always acquires the same image of the target region 70. 11A and 11B, the target area 70 is abstractly shown by an ellipse. By binarizing the difference image (or the ratio image, the same applies hereinafter) of both images, the first dark defect candidate image 821 indicating the first dark defect candidate region 721 shown in FIG. 11C is acquired. Through the above process, a region in the target region 70 where the captured image 811 is darker than the reference image 812 and satisfies a predetermined condition is acquired as the first dark defect candidate region 721. In the present embodiment, the pixel value is “1” in the first dark defect candidate area 721, and the pixel value is “0” in the other areas.

  When the first dark defect candidate region 721 is acquired, the area filter unit 535 deletes the first dark defect candidate region 721 whose area is smaller than a predetermined value, and the remaining first dark defect candidate region 721 is removed. The image shown is acquired as the final first dark defect candidate image 821 (more precisely, the first dark defect candidate data 921 which is the image data showing the first dark defect candidate region 721).

  The plurality of captured images acquired by the image capturing unit 3 are sequentially selected as the processing target, and the first dark defect candidate images 821 are acquired in the number equal to the number of captured images (step S14).

  As shown in FIG. 6, in the second dark defect candidate acquisition unit 522, the captured image data 911 and the reference image data 912 are input to the expansion processing unit 541 from the pre-alignment unit 532, and expansion processing is performed on the captured image and the reference image. Be seen. The expansion process here is a process of expanding a bright region in the multi-valued image. This causes the small dark areas to disappear. The data of the captured image and the reference image are further input to the contraction processing unit 542, and contraction processing is performed on the captured image and the reference image. The contraction process here is a process of contracting a bright area in the multi-valued image. As a result, the bright area is returned to almost the original size. As a result, the large dark areas in the original captured image and the reference image are maintained in their original states, and the small dark areas disappear.

  FIG. 12A is a diagram illustrating an image 813 after the expansion and contraction processing is performed on the captured image 811 of FIG. 11A. FIG. 12B is a diagram illustrating an image 814 after the reference image 812 of FIG. 11B has been subjected to expansion and contraction processing. In both images, small dark areas of the original image disappear. The comparison unit 543 generates difference image data of these images. The difference image is binarized by the binarization unit 544 using the second reference value. The process of generating the difference image and binarizing it is the same as that of the shake comparing unit 533 and the binarizing unit 534 in FIG. 5 except that the shake is not performed. The same applies to the point that a ratio image may be used instead of the difference image. As a result, as shown in FIG. 12C, an area in the target area 70 where the captured image 811 is darker than the reference image 812 and which satisfies a predetermined condition is acquired as the second dark defect candidate area 722. . In the present embodiment, the pixel value is “1” in the second dark defect candidate area 722, and the pixel value is “0” in the other areas.

  When the second dark defect candidate region 722 is acquired, the area filter unit 545 deletes the second dark defect candidate region 722 whose area is smaller than a predetermined value, and the remaining second dark defect candidate region 722 is removed. The image shown is acquired as the final second dark defect candidate image 822 (more precisely, the second dark defect candidate data 922 which is image data showing the second dark defect candidate region 722).

  The plurality of picked-up images acquired by the image pickup unit 3 are sequentially selected as the processing target, and the second dark defect candidate images 822 in the number equal to the number of picked-up images are acquired (step S15).

  As shown in FIG. 8, the first dark defect candidate data 921 and the second dark defect candidate data 922 generated from the same captured image and the corresponding reference image are input to the logical product calculation unit 561. The data of the logical product image obtained from each combination of the first dark defect candidate image 821 and the second dark defect candidate image 822 is input to the region selection unit 562. Each logical product image is an image showing a dark defect candidate region generated from the first dark defect candidate region and the second dark defect candidate region. FIG. 13: is a figure which illustrates the dark defect candidate image 824 which shows the dark defect candidate area | region 724 produced | generated from the 1st dark defect candidate area | region 721 of FIG. 11C and the 2nd dark defect candidate area | region 722 of FIG. 12C.

  Since the first dark defect candidate region is obtained from the difference image between these images after the captured image and the reference image have been aligned with each other, the presence of a defect is highly reliable. However, the small first dark defect candidate area may have occurred due to some cause such as noise. On the other hand, since the second dark defect candidate area 722 is obtained from these difference images after performing the small area removal processing for deleting the small dark area on the captured image and the reference image, it is possible to detect a small false defect. It is not very popular. However, due to the processing time, there is a possibility that a false defect may be detected due to the incorrect alignment. Therefore, in the present embodiment, by obtaining a logical product image of the first dark defect candidate image and the second dark defect candidate image, a dark defect candidate image showing a dark defect candidate region with improved reliability is obtained. .

  In the area selection unit 562, when a plurality of dark defect candidate images are overlapped, the area where a predetermined number or more of dark defect candidate areas overlap is maintained as a dark defect candidate area. Here, the “overlapping area” may be a logical sum or a logical product of a plurality of overlapping areas. In the present embodiment, when the number of overlapping areas is 2 or more, it is maintained as a dark defect candidate area. The area selecting unit 562 acquires one dark defect candidate image showing the limited dark defect candidate area from the plurality of dark defect candidate images showing the dark defect candidate area (step S16). Through the above process, a dark region with respect to the reference image is acquired as a dark defect candidate region from the plurality of captured images acquired in the plurality of illumination states while referring to the corresponding reference image.

  Since the dark defect candidate area is limited by the area selection unit 562, it is possible to actually use a logical sum calculation unit instead of the logical product calculation unit 561. In any case, the reliability of the dark defect candidate region can be improved by acquiring the dark defect candidate region based on the first dark defect candidate region and the second dark defect candidate region. The number of overlapping areas when maintained as the dark defect candidate area may be three or more, or may be one or more. When the number of overlapping regions is set to 1 or more, all dark defect candidate regions are maintained, and the region selection unit 562 simply generates a logical sum image of a plurality of dark defect candidate images as one dark defect candidate image. To do.

  On the other hand, in the second bright defect candidate acquisition unit 524, as shown in FIG. 7, the data of the light-dark reverse region image showing the light-dark reverse region is obtained by the process in which the light-dark is replaced with that of the second dark defect candidate acquisition unit 522. To get.

  Specifically, in the second bright defect candidate acquisition unit 524, the captured image data 911 and the reference image data 912 are input from the pre-alignment unit 532 to the contraction processing unit 551, and contraction processing is performed on the captured image and the reference image. The contraction process here is a process of contracting a bright region in a multi-valued image, and is also an expansion process for a dark region. This causes the small bright areas to disappear. The data of the captured image and the reference image are further input to the expansion processing unit 552, and the expansion process is performed on the captured image and the reference image. The expansion process here is a process of expanding a bright region in a multi-valued image and a process of contracting a dark region. As a result, the dark area is returned to its original size. As a result, the large bright areas in the original captured image and the reference image are maintained in their original states, and the small bright areas disappear.

  The comparison unit 553 generates difference image data of these images. The difference image is binarized by the binarization unit 554 using the third reference value. A ratio image may be used instead of a difference image. In the present embodiment, the pixel value is "1" in the light-dark reverse region, and the pixel value is "0" in the other regions.

  When the bright / dark reverse region is acquired, the area filter unit 555 deletes the bright / dark reverse region having an area smaller than a predetermined value, and the image showing the remaining bright / dark reverse region is a final bright / dark reverse region image (exactly Is acquired as bright / dark reverse area data 923) which is image data showing the bright / dark reverse area.

  The plurality of picked-up images acquired by the image pickup unit 3 are sequentially selected as a processing target, and the number of bright / dark reverse region images equal to the number of picked-up images is acquired (step S17). Through the above processing, a bright region with respect to the reference image is acquired as a bright / dark reverse region from each of the plurality of captured images acquired in the plurality of illumination states while referring to the corresponding reference image.

  As shown in FIG. 9, the plurality of bright / dark reverse area data 923 are input to the logical sum operation unit 563, and data of the logical sum image of the bright / dark reverse area image is generated. 14A and 14B are diagrams illustrating a bright / dark reverse region image 823 showing a bright / dark reverse region 723. FIG. 14C shows a logical sum image 825 of these images. The actual number of bright / dark reverse region images is equal to or less than the number of captured images acquired by one image capturing unit 3, and preferably 2 or more.

  The data of the logical sum image is input to the candidate limiting unit 564. The candidate limiting unit 564 excludes, from the defect candidates, dark defect candidate regions that do not overlap any of the bright and dark reverse regions in the logical sum image among the dark defect candidate regions. This further limits the dark defect candidate area. FIG. 15 is a diagram showing the dark defect candidate image 824 of FIG. 13 and the logical sum image 825 of FIG. 14C in an overlapping manner. The dark defect candidate area 724 on the right side is maintained as a defect candidate because it overlaps with the bright / dark inverse area 723. The left dark defect candidate region 724 does not overlap any of the bright / dark reverse regions 723 and is therefore excluded from the defect candidates (step S18).

  It should be noted that the dark defect candidate image 824 may be sequentially overlapped with each of the light / dark reverse region images 823 without obtaining the logical sum image, and the presence or absence of the overlap between the dark defect candidate region 724 and the light / dark reverse region 723 may be confirmed.

  Only the overlap of the dark defect candidate region 724 and the bright / dark reverse region 723 which is equal to or larger than a predetermined area may be detected as the presence of the overlap. In this case, the dark defect candidate region 724 whose overlapping with the bright / dark reverse region 723 is less than a predetermined area is excluded from the defect candidates. Further, when the overlap between the dark defect candidate area 724 and the bright / dark inverse area 723 is less than a predetermined ratio with respect to the area of the dark defect candidate area 724, the dark defect candidate area 724 is excluded from the defect candidates. Good. In this way, among the dark defect candidate regions, those that do not overlap any of the bright and dark reverse regions by a predetermined condition or more are excluded from the defect candidates. In other words, among the dark defect candidate regions, one that overlaps with any one of the bright and dark reverse regions by the predetermined condition or more is maintained as the dark defect candidate.

  After that, a final dark defect region is acquired based on the dark defect candidate region by a process such as deleting a dark defect candidate region having a predetermined area or less as necessary. That is, the presence of a dark defect is acquired (step S19). By the above processing, the position of the defect when the defect exists in the target area 70 viewed from one image pickup unit 3 is detected.

  One captured image is displayed on the display unit of the computer 12, and a dark defect area is displayed on the target area 70.

  When a concave or convex defect is present on the surface of the target object 9, the defect appears bright or dark in the captured image when polarized light is irradiated. However, if an attempt is made to detect a defect using only a dark area, surface stains due to unnecessary oil or paint will also be detected as a defect. Therefore, by using the light-dark reverse region image, the dark region due to the surface color is removed as noise, and the overdetection of defects is suppressed. In particular, since the appearance of light and dark due to the presence of a defect is complicated, by using a plurality of light and dark reverse region images, it is possible to remove with high accuracy a defect caused by the surface color.

  The minimum number of captured images acquired by one image capturing unit 3 is 2, but preferably 3 or more. That is, the light irradiation unit 4 can be set to three or more illumination states different from each other, for example, the target object 9 can be irradiated with light from three or more directions, and the imaging control unit 51 controls the light irradiation unit to have three or more illumination states. The imaging unit 3 acquires an image during the illumination state. When the preferable illumination state is known, at least one of the three or more captured images acquired based on the information may be selected as the captured image of the processing target. By preparing three or more captured images, more appropriate defect detection can be easily performed.

  As described above, the defect detection device 1 also has a function of detecting bright defects. The detection of the bright defect is the same as steps S14 to S19 except that the bright and dark are reversed from the detection of the dark defect. The configuration of the first bright defect candidate acquisition unit 523 is the same as the first dark defect candidate acquisition unit 521 of FIG. 5 except that the calculation order and the reference value are different. That is, the first bright defect candidate acquisition unit 523 aligns the captured image and the reference image, obtains a difference image, and binarizes the difference image. As a result, a first bright defect candidate image indicating the first bright defect candidate area is acquired. The second bright defect candidate acquisition unit 524 performs contraction and expansion processing on the captured image and the reference image to acquire a difference image, and binarizes the difference image. As a result, the second bright defect candidate image showing the second bright defect candidate area is acquired.

  With the configuration shown in FIG. 8, a bright defect candidate image indicating a bright defect candidate area is generated from a plurality of sets of the first bright defect candidate image and the second bright defect candidate image. On the other hand, the second dark defect candidate acquisition unit 522 functions as a bright / dark reverse region acquisition unit and acquires the second dark defect candidate image as a bright / dark reverse region image. Then, a logical sum image of a plurality of light and dark reverse region images is generated.

  Of the bright defect candidate areas, those that do not overlap any of the bright and dark reverse areas by a predetermined condition or more are excluded from the defect candidates, and the bright defect candidate areas are limited. Then, a bright defect image indicating the presence or absence of a bright defect is generated using the bright defect candidate image. By the above processing, it is possible to prevent the bright defect from being excessively detected due to stains of bright color.

  The dark defect area and the bright defect area are displayed on the display as areas with different colors on one captured image, for example.

  When dark defect detection and bright defect detection are performed, the second bright defect candidate used for light defect detection is limited in order to limit the dark defect candidate area from the viewpoint of improving the processing speed by reducing the amount of information. It is preferable to use the region as a light / dark inverted region. In order to limit the bright defect candidate area, it is preferable to use the second dark defect candidate area used for detecting the dark defect as the light / dark reverse area. That is, it is preferable that the step of acquiring the bright / dark reverse area when acquiring one of the dark defect candidate area and the bright defect candidate area is included in the step of acquiring the other of the dark defect candidate area and the bright defect candidate area. However, if the calculation time is allowed, it is conceivable to use a more preferable bright / dark reverse region.

  The reference value, which is a threshold value for binarization when obtaining the dark defect candidate region or the bright defect candidate region, is set to an optimum value that suppresses overdetection while suppressing overdetection. On the other hand, the bright / dark reverse region is used to suppress the detection of false defects caused by surface color such as dirt and noise. Therefore, by setting the reference value, which is a threshold value for binarization when obtaining the light-dark reverse region, to a value that makes the light-dark reverse region even a doubtable light-dark reverse region, the true defect is excluded as a stain or the like. Is suppressed.

  Therefore, the reference value (the third reference value in the above description) used by the second bright defect candidate acquisition unit 524 when acquiring the dark defect candidate area is set to the second bright defect candidate when acquiring the bright defect candidate area. By making the value smaller than the reference value used in the acquisition unit 524, it is possible to obtain a more accurate dark defect candidate area. Even when acquiring the bright defect candidate area, the reference value used by the second dark defect candidate acquiring section 522 is set to the reference value used by the second dark defect candidate acquiring section 522 when acquiring the dark defect candidate area. By making the value smaller than (the second reference value in the above description), a more preferable bright defect candidate area can be obtained.

  As described above, the reference value for binarization when acquiring the bright / dark reverse region may be different from the reference value for binarization when acquiring the second dark defect candidate region or the second bright defect candidate region, and is preferable. The reference value for binarization when acquiring the bright / dark inverse region is looser than the reference value for binarization when acquiring the second dark defect candidate region or the second bright defect candidate region. The area of the bright / dark reverse region is larger than the area of the second dark defect candidate region or the second bright defect candidate region.

  In the above description, the detection of the dark defect and the detection of the bright defect are distinguished and described, but the defect candidate in the defect detection apparatus 1 is one or both of the dark defect candidate and the bright defect candidate. Generally, the dark defect candidate and the bright defect candidate in the above description can be expressed as “defect candidates”, and the dark defect candidate region and the bright defect candidate region can be expressed as “defect candidate regions”. The defect candidate image and the bright defect candidate image can be expressed as “defect candidate image”. The first dark defect candidate region and the first bright defect candidate region can be expressed as “first defect candidate regions”, and the second dark defect candidate region and the second bright defect candidate region are referred to as “second defect candidate regions”. Can be expressed.

In the above-described embodiment, the defect acquisition unit 52 acquires one of a dark region and a bright region with respect to the reference image as a defect candidate region from the captured image used for defect detection while referring to the corresponding reference image. . Further, the defect acquisition unit 52 acquires the other of the dark region and the bright region with respect to the reference image as another defect candidate region from the captured image used for the defect detection while referring to the corresponding reference image. Then, in the case of efficiently performing the processing, the light-dark reverse region corresponding to the defect candidate region is acquired when acquiring the other defect candidate region. Further, in the case of improving the defect detection accuracy without being concerned with the processing efficiency, the bright / dark reverse region corresponding to the defect candidate region is acquired by using a reference value different from the reference value in the acquisition of the other defect candidate region. To be done.

  Various modifications can be made to the defect detection device 1.

  The arrangement and number of the light source units 4a, 4b, 4c and the image pickup unit 3 may be changed as appropriate. The illumination state by the light irradiation unit 4 may be variously changed. Two of the plurality of light source units may be turned on, or three of the light source units may be turned on. In the light irradiation unit 4, the lighting state may be changed by moving the light source unit.

  When the method of acquiring the first defect candidate area is expressed as the first method and the method of acquiring the second defect candidate area is expressed as the second method in the above embodiment, the first method and the second method Various methods can be used as long as they are different from the above method. Thereby, over-detection or omission of detection can be efficiently suppressed.

  Alternatively, the defect candidate area may be obtained by only one method. For example, without determining the second defect candidate area in the above-described embodiment, the area selecting unit 562 determines that one of the first defect candidate areas of the plurality of first defect candidate images that overlaps a predetermined number or more is a defect candidate area. It may be acquired. Alternatively, without determining the first defect candidate area in the above-described embodiment, the area selecting unit 562 determines, as the defect candidate area, the second defect candidate areas of the plurality of second defect candidate images that overlap by a predetermined number or more. It may be acquired.

In the above embodiment, only one captured image may be used for defect detection. The number of captured images used when acquiring the bright / dark reverse region image is preferably two or more. In order to perform imaging efficiently, it is preferable that at least one captured image used for defect detection be included in a plurality of captured images acquired in a plurality of illumination states to acquire a bright / dark reverse region.

  The computer 12 may be dedicated hardware, or partially dedicated hardware may be used. When performing visual inspection at high speed, it is preferable that parallel processing be performed by a computer or dedicated hardware.

  As long as substantially the same processing is performed, the processing order can be changed appropriately. The processing order described in the above embodiment is merely an example. For example, the step of acquiring the first dark defect candidate area, the step of acquiring the second dark defect candidate area, and the step of acquiring the bright / dark reverse area may be performed in any order or in parallel. The process of detecting a dark defect and the process of detecting a bright defect may be performed in any order or in parallel. The defect candidate area may be directly treated as a defect area.

  The reference image may be generated from the captured image. That is, the defect candidate area may be acquired by so-called self-comparison. For example, the reference image in which the dark defect disappears is acquired by performing the process of expanding the bright region and then the process of contracting the captured image. By performing the process of contracting the bright region on the captured image and then the process of expanding the bright region, the reference image in which the bright defect disappears is acquired.

  The defect detection device 1 may be used for detecting defects on the surface of another object such as various substrates and films on which patterns are formed. The defect detection apparatus 1 is particularly suitable for inspecting an object that is likely to be overdetected by having a satin-shaped area (not limited to a metal surface) on the surface.

  The configurations of the above-described embodiment and each modification may be appropriately combined unless they contradict each other.

DESCRIPTION OF SYMBOLS 1 Defect detection device 3a, 3b, 3c Imaging part 4 Light irradiation part 9 Object 12 Computer 51 Imaging control part 52 Defect acquisition part 70 Target area 80 Program 721 1st dark defect candidate area 722 2nd dark defect candidate area 723 Bright reverse Area 724 Dark defect candidate area 811 Captured image 812 Reference image S11 to S19 Step

Claims (9)

A defect detection device for detecting defects on the surface of an object,
In a plurality of different biased illumination states, a light irradiation unit capable of irradiating an object with light,
An image capturing unit that acquires an image of a target region on the surface of the target object as a captured image,
An illumination control by the light irradiation unit, and an imaging control unit that controls acquisition of an image by the imaging unit,
A dark region and a bright region with respect to the reference image while referring to a corresponding reference image from at least one captured image acquired in at least one illumination state included in the plurality of illumination states and used for defect detection. get one of the defect candidate regions, from each of the plurality of captured images acquired by two or more lighting states included in the plurality of lighting conditions, with respect to the reference image while referring to the reference image corresponding Obtaining the other of the dark area and the bright area as the light-dark reverse area, among the defect candidate areas, after excluding those that do not overlap any light-dark reverse area by a predetermined condition or more from the defect candidates, based on the defect candidate area. A defect acquisition unit that acquires the existence of defects,
A defect detection device comprising: The defect detection device according to claim 1, wherein
The defect detecting section, a plurality of captured images obtained by the plurality of lighting conditions, the defect detecting apparatus characterized by sequentially selecting as IMAGING images that are used for defect detection. The defect detection device according to claim 1 or 2, wherein
The at least one captured image used for defect detection includes a first captured image and a second captured image,
The defect acquisition unit acquires, from the first captured image, one of a dark region and a bright region with respect to the first reference image as the defect candidate region while referring to the corresponding first reference image, and the second obtained from the captured image, the other dark regions and bright regions as another defect candidate region with respect to the second reference picture with reference to the corresponding second reference image,
When acquiring the bright dark reverse area that corresponds to the prior SL defect candidate regions, wherein the other defect candidate area from the second image is acquired, a dark reverse area corresponding to the other defect candidate regions The defect detection device, wherein the defect candidate area is acquired from the first captured image when the defect is detected. The defect detection device according to any one of claims 1 to 3,
The defect acquisition unit acquires a first defect candidate area by a first method from the at least one captured image used for defect detection, and a second method different from the first method by a second method. A defect detecting apparatus, which acquires two defect candidate areas and acquires the defect candidate areas based on the first defect candidate area and the second defect candidate area. The defect detection device according to claim 4,
The first method is a method of acquiring the first defect candidate area from a difference image of these images after performing alignment between the captured image and the reference image,
The second method is a method of performing the small area removal processing on the captured image and the reference image, and then acquiring the second defect candidate area from the difference image of these images. . A defect detection method for detecting defects on the surface of an object,
a) While the object is being illuminated with light in each of a plurality of differently biased illumination states, the imaging unit acquires an image of the target region on the surface of the object to obtain a plurality of captured images. The process of acquiring
b) a dark region with respect to the reference image while referring to a corresponding reference image from at least one captured image acquired in at least one illumination state included in the plurality of illumination states and used for defect detection; Acquiring one of the bright areas as a defect candidate area,
c) From each of the plurality of captured images acquired in two or more illumination states included in the plurality of illumination states, refer to the corresponding reference image and select the other of the dark region and the bright region with respect to the reference image. A step of acquiring as a light and dark reverse area,
d) Among the defect candidate areas, a step of excluding, from the defect candidates, areas that do not overlap any of the light and dark reverse areas by a predetermined condition or more, and acquiring the existence of a defect based on the defect candidate areas,
A defect detection method comprising: The defect detection method according to claim 6,
Defects each multiple captured image acquired by the previous SL plurality of lighting conditions, while sequentially selecting the captured images used for defect detection, characterized in that the b) to d) steps are repeated Detection method. The defect detection method according to claim 6 or 7, wherein
The at least one captured image used for defect detection includes a first captured image and a second captured image,
In the step b), one of a dark region and a bright region with respect to the first reference image is acquired as the defect candidate region from the first captured image while referring to the corresponding first reference image,
The defect detection method,
from e) said second image, and acquires the other dark area and bright area relative to the second reference picture with reference to the corresponding second reference image as another defect candidate regions step,
Further equipped with,
When acquiring the light / dark reverse region corresponding to the defect candidate region, the other defect candidate region is acquired from the second captured image, and when acquiring the light / dark reverse region corresponding to the other defect candidate region, The defect detection method, wherein the defect candidate area is acquired from the first captured image . A program for causing a computer to detect a defect in the target region from a plurality of images of the target region on the surface of the target, wherein the computer executes the program,
a) preparing a plurality of captured images of the target region and corresponding plurality of reference images acquired in a plurality of differently biased illumination states,
b) a dark region with respect to the reference image while referring to a corresponding reference image from at least one captured image acquired in at least one illumination state included in the plurality of illumination states and used for defect detection; Acquiring one of the bright areas as a defect candidate area,
c) From each of the plurality of captured images acquired in two or more illumination states included in the plurality of illumination states, refer to the corresponding reference image and select the other of the dark region and the bright region with respect to the reference image. A step of acquiring as a light and dark reverse area,
d) Among the defect candidate areas, a step of excluding, from the defect candidates, areas that do not overlap any of the light and dark reverse areas by a predetermined condition or more, and acquiring the presence of a defect based on the defect candidate areas,
A program characterized by causing to execute.

Images