JP2019045451A - Inspection apparatus, inspection method and program - Google Patents

Abstract

To provide an inspection apparatus and the like that improve convenience of inspection.SOLUTION: An inspection apparatus that detects a defect of an inspection object by comparing an inspection image 200 which is obtained by imaging the inspection object with a reference image 100 calculates a difference of pixel values between the inspection image 200 and the reference image 100 to create a difference image 300, and creates two mask images 301a and 301b using edge expansion images 202 and 102 in which edges of the inspection image 200 and the reference image 100 are expanded relative to the difference image 300. Further, noise removal is performed by creating a composite image 302 in which two mask images 301a and 301b are synthesized, subtracting a predetermined value from a pixel value of the composite image 302, and making the pixel value of a pixel in which the pixel value is smaller than 0 as 0. The inspection apparatus detects the defect using the composite image 302 after performing the noise removal.SELECTED DRAWING: Figure 5

Description

  The present invention relates to an inspection apparatus, an inspection method, and a program using image processing technology.

  At a manufacturing site, it is carried out to inspect the presence or absence of a defect of a product from an inspection image obtained by photographing the product with a photographing device. As the inspection method, defects are often detected by comparing an inspection image with a reference image.

  In Patent Document 1, a defect image of dirt and omission is created using a difference image obtained by taking the difference between the pixel value of the reference image and the pixel value of the inspection image, and the pixel value of the defect image is binary with a predetermined threshold. It is described that the defect candidate of dirt or omission is extracted and the area is used to determine whether the defect candidate is a true defect or not.

Patent No. 4844784

  However, the above inspection method has problems in terms of the convenience of the inspection. For example, in the conventional method, the processing speed is reduced because the comparison processing with the threshold value is performed on the pixel values of all pixels of the defect image. Further, since noise is generated in the defect image due to the variation of the sensitivity of the light receiving element of the imaging device, etc., it is necessary to carefully set the threshold considering noise in the subsequent defect candidate extraction processing.

  The present invention has been made in view of the above problems, and an object of the present invention is to provide an inspection apparatus and the like in which the convenience of the inspection is improved.

  A first invention for solving the above-mentioned problems is an inspection apparatus for detecting a defect of the inspection object by comparing an inspection image obtained by photographing the inspection object with a reference image, the inspection image And a difference image generation unit that generates a difference image by taking a difference of pixel values between the reference image and the reference image, and subtracting a predetermined value from pixel values of the difference image or an image generated based on the difference image; Detection of defects using a noise removal unit that performs noise removal by setting the pixel value of a pixel whose value is less than 0 to 0, an image after the noise removal is performed, or an image created based on the image And a defect detection unit that performs the inspection.

  In the present invention, at the time of detection of a defect from an image, noise removal by subtracting a predetermined value from a pixel value has already been performed, so it is not necessary to consider noise when detecting a defect. Further, the part where the pixel value is 0 at the time of noise removal does not have to be considered at the time of detection of a defect, and can be excluded from the analysis target, so that the processing speed can be increased. Therefore, the convenience of the inspection can be improved.

The defect detection unit extracts a defect candidate by binarizing the pixel value of the image after the noise removal or the image created based on the image with a predetermined threshold value, and based on the area of the defect candidate It is desirable to determine whether the defect candidate is a true defect.
At the time of detection of a defect, for example, the pixel value of the image after noise removal can be compared with a threshold value to extract a defect candidate. At this time, noise may not be taken into consideration when setting the threshold as described above, and since the portion where the pixel value is 0 can be excluded from the analysis target, it is necessary to perform comparison and determination processing of the pixel value and the threshold over the entire image Processing speed can be realized.

An edge expansion image obtained by expanding an edge of the inspection image is used to perform a mask process for removing an edge portion generated in the difference image to create a first mask image, and expand an edge of the reference image. The image processing apparatus further includes a mask processing unit that performs mask processing for removing an edge portion generated in the difference image using an edge expansion image to create a second mask image, and the noise removal unit includes the first and second noise removal units. It is desirable to perform the noise removal by subtracting the predetermined value from the pixel value of the combined image obtained by combining the mask images of
By performing the mask processing described above, even when an edge portion appears in a difference image due to positional deviation between an inspection image and a reference image, it is possible to prevent erroneous detection as a defect, and Defects can be reliably detected by using a composite image obtained by combining two mask images for defect detection. In the present invention, the noise removal process is performed only once by performing the noise removal on the composite image, thereby speeding up the process.

  A second invention is an inspection method for detecting a defect of the inspection object by comparing an inspection image obtained by photographing the inspection object with a reference image, and a computer is configured to compare the inspection image and the reference image. And subtracting a predetermined value from the pixel value of the image created based on the difference image or the difference image, and calculating the difference between the pixel values between the pixels, and subtracting the predetermined value from the pixel value of the image created based on the difference image Performing a step of removing noise by setting the pixel value to 0, and a step of detecting a defect using the image after the removal of the noise or an image created based on the image It is an inspection method to be characterized.

  A third invention is an inspection apparatus for detecting a defect of the inspection object by comparing a computer with an inspection image obtained by photographing the inspection object with a reference image, and the inspection image and the reference image A predetermined value is subtracted from the pixel value of an image created based on the difference image or the difference image or the difference image, and a pixel value becomes less than 0. A noise removal unit that performs noise removal by setting the pixel value of each pixel to 0, and a defect detection unit that performs defect detection using an image after the noise removal or an image created based on the image And a program for functioning as an inspection device.

  According to the present invention, it is possible to provide an inspection apparatus etc. in which the convenience of inspection is improved.

FIG. 1 shows an inspection system 1; FIG. 2 is a diagram showing a hardware configuration of an inspection device 3. The example of the reference | standard image 100 and the test | inspection image 200. FIG. FIG. 2 is a diagram showing a functional configuration of the inspection device 3. The flowchart which shows the flow of the inspection method. The example of the edge image 101 and the edge expansion image 102. FIG. The example of the edge image 201 and the edge expansion image 202. FIG. The example of the difference image 300. The example of mask image 301a, 301b. An example of a composite image 302. The figure explaining a noise removal process and the extraction process of a defect candidate. The example of the difference image 400. The example of mask image 401a, 401b. An example of a composite image 402.

  Hereinafter, preferred embodiments of the present invention will be described in detail based on the drawings.

(1. Inspection system 1)
FIG. 1 is a view showing an inspection system 1 including an inspection apparatus 3 according to an embodiment of the present invention. As shown in FIG. 1, the inspection system 1 includes an inspection device 3, an imaging device 5, and the like.

  The inspection device 3 controls the entire inspection in the inspection system 1 and performs defect inspection of an inspection object according to a procedure described later. The inspection target is, for example, a pattern (printed matter) printed on the sheet 10, and it is checked whether there is a defect in the pattern based on the inspection image obtained by photographing the pattern. However, the inspection target is not limited to this, and may be a circuit pattern or a printing plate.

  FIG. 2 is a diagram showing a hardware configuration of the inspection apparatus 3. As shown in FIG. 2, the inspection apparatus 3 can be realized by a computer configured by connecting, for example, the control unit 31, the storage unit 32, the input unit 33, the display unit 34, and the communication control unit 35 via the bus 36. However, without being limited to this, various configurations can be taken as appropriate.

  The control unit 31 includes a CPU, a ROM, a RAM, and the like. The CPU calls a program related to the processing of the inspection apparatus 3 stored in the storage unit 32 or a storage medium such as a ROM to a work area on the RAM and executes the program. The ROM is a non-volatile memory, which permanently holds a boot program, a program such as a BIOS, data and the like. The RAM is a volatile memory, and temporarily stores programs and data loaded from the storage unit 32 and the ROM, and includes a work area used by the control unit 31 to perform various processes.

  The storage unit 32 is a hard disk drive, a solid state drive, a flash memory or the like, and stores a program executed by the control unit 31, data necessary for program execution, an OS, and the like. These programs and data are read and executed by the control unit 31 as needed.

The input unit 33 is used to input various settings to the inspection apparatus 3.
The display unit 34 is, for example, a liquid crystal display.
The communication control unit 35 is a communication interface that mediates communication via a network or the like, and communicates with the imaging device 5 or the like.
The bus 36 is a path that mediates the exchange of control signals, data signals, and the like between the respective units of the inspection apparatus 3.

  The imaging device 5 captures an inspection object. The photographing device 5 includes a plurality of light receiving elements, a condensing lens, and the like, and converts the light received by each light receiving element into an electric signal. For example, an area camera such as a CCD (Charge Coupled Device) camera is used as the photographing device 5, but the invention is not limited thereto, and a line camera or the like may be used. Although not particularly illustrated, in the present embodiment, illumination for photographing an inspection object is separately provided.

  In the present embodiment, when performing a defect inspection by the inspection device 3, a reference image used for the inspection is stored in the storage unit 32. FIG. 3A is an example of this reference image 100. FIG. The reference image 100 is a positive image representing a defect-free inspection object, and can be obtained, for example, by imaging a defect-free inspection object.

  On the other hand, the inspection apparatus 3 detects a defect on the inspection image obtained by photographing the inspection object by the imaging apparatus 5 in comparison with the reference image 100. FIG. 3B shows an example of the inspection image 200. In the present embodiment, it is assumed that there are a dirt defect a and a drop-out defect b in the illustrated portion. In the present embodiment, the reference image 100 and the inspection image 200 are gray scale images, and the contrast is represented by pixel values (gradation values). The pixel value increases as the brightness increases.

  The dirt defect a is a portion where a change in color (a concept including brightness) from the inspection object due to dirt occurs in the image, and the omission defect b is an inspection due to the fact that a portion to be originally formed in the inspection object It is the part where a change in color from the subject occurs in the image.

  In the present embodiment, it is assumed that patterns (patterns) to be inspected are formed by arranging high-brightness and low-brightness backgrounds and forming low-brightness and high-brightness toroidal portions on the respective backgrounds. In addition, it is assumed that the luminance of the portion where the luminance is originally high decreases due to the dirt defect a, and the luminance of the portion where the luminance is originally low becomes high due to the omission defect b.

  FIG. 4 is a diagram showing a functional configuration of the inspection apparatus 3. As shown in FIG. 4, the inspection apparatus 3 includes a difference image generation unit 11, a mask processing unit 12, an image synthesis unit 13, a noise removal unit 14, a defect detection unit 15, and the like.

  The difference image generation unit 11 generates a difference image by calculating the difference between the pixel values of the reference image 100 and the inspection image 200. In this embodiment, in order to separately detect the above-described dirt defect a and dropout defect b, a difference image obtained by subtracting the pixel value of the inspection image 200 from the pixel value of the reference image 100 and the reference image from the pixel values of the inspection image 200 The difference image which subtracted 100 pixel values is each created.

  The mask processing unit 12 performs mask processing on each difference image using an edge expansion image obtained by expanding the edges of the reference image 100 and the inspection image 200 to create two mask images.

  The image combining unit 13 combines the two mask images to create a combined image.

  The noise removing unit 14 performs noise removal by subtracting a predetermined value from the pixel value of the composite image and setting the pixel value of the pixel whose pixel value is less than 0 to 0.

  The defect detection unit 15 detects the dirt defect a and the missing defect b using the composite image after the noise removal.

(2. Inspection method)
Next, the inspection method according to the present embodiment will be described. FIG. 5 is a flowchart showing the flow of the inspection method, and each step of FIG. 5 is executed by the control unit 31 of the inspection apparatus 3. Here, as described above, the reference image 100 used for the inspection is stored in the storage unit 32 of the inspection device 3, and the inspection image 200 obtained by photographing the inspection object by the imaging device 5 is acquired by the control unit 31 of the inspection device 3 Be in the

  In the present embodiment, first, the inspection device 3 extracts an edge of the reference image 100 (S1), and creates an edge expansion image 102 in which the edge of the reference image 100 is expanded (S2). FIG. 6A shows an example of the edge image 101 in which the edge of the reference image 100 is extracted, and FIG. 6B shows an example of the edge expansion image 102 in which the edge is expanded. A known method can be applied to edge extraction and expansion, and thus detailed description will be omitted.

  The inspection apparatus 3 also performs edge extraction (S3) and edge expansion on the inspection image 200 obtained by photographing the inspection object as described above to create an edge expansion image 202 (S4). FIG. 7A is an example of an edge image 201 obtained by extracting an edge from the inspection image 200, and FIG. 7B is an example of an edge expanded image 202 obtained by expanding the edge.

  In the present embodiment, using the reference image 100, the inspection image 200, and the edge expansion images 102 and 202, the inspection of the dirt defect a and the omission defect b in the inspection image 200 is performed.

  As the inspection of the stain defect a, first, a difference image 300 is created by subtracting the pixel value of the inspection image 200 from the pixel value of the reference image 100 (S5). FIG. 8 is an example of the difference image 300. Here, calculation is performed by subtracting the pixel value of the pixel of the inspection image 200 from the pixel value of the pixel of the reference image 100 between the pixels at corresponding positions of the reference image 100 and the inspection image 200. If the value after calculation is less than 0, the pixel value of that pixel is set to 0.

  By subtracting the pixel value of the inspection image 200 from the pixel value of the reference image 100, the stain defect a appears in the difference image 300. However, in the example of FIG. Are slightly misaligned with respect to the reference image 100, and part of the edge portion of the pattern is also apparent. Since this is not a defect, it is necessary to remove it from the difference image 300 in the subsequent processing.

  Therefore, the inspection apparatus 3 performs mask processing on the difference image 300 using the edge expansion image 202, and creates a mask image 301a (first mask image) (S6). FIG. 9A shows an example of the mask image 301a. Here, the pixel value of the pixel of the difference image 300 is compared with the pixel value of the pixel at the corresponding position in the edge expansion image 202, and if the pixel value of the difference image 300 is higher, the pixel value of that pixel is used as it is. The pixel value of that pixel is set to 0 otherwise.

  The inspection apparatus 3 further mask-processes the difference image 300 using the edge expansion image 102, and creates a mask image 301b (second mask image) (S7). FIG. 9B is an example of the mask image 301 b. The mask processing method is the same as described above.

  As shown in FIGS. 9A and 9B, by performing mask processing with the edge expansion images 102 and 202, the edge portions described above are removed. However, the appearance (remaining manner) of the dirt defect a is different depending on which of the edge expansion images 102 and 202 is used to perform the mask processing, so in the present embodiment, the dirt image a is detected with certainty. , 301b are synthesized to generate a synthesized image 302 (S8).

  FIG. 10 shows an example of the composite image 302. Here, the pixel values of the pixels at corresponding positions are compared between the mask images 301a and 301b, and if the pixel values of the pixels of the mask image 301a are higher, the pixel values of those positions are the pixel values of the pixels of the mask image 301a. When the pixel value of the pixel of the mask image 301b is higher than the pixel value, the pixel value of the position is set as the pixel value of the pixel of the mask image 301b. If the pixel values of the pixels of the mask images 301a and 301b are the same, the pixel value is set as the pixel value of that position.

  The inspection apparatus 3 performs noise removal by subtracting a predetermined value from the pixel value of each pixel and setting the pixel value of the pixel whose pixel value is less than 0 to 0 in the composite image 302 (S9).

  That is, in the present embodiment, noise (a slight change in pixel value) is generated in the pixel value of the composite image 302 due to variations in sensitivity of light receiving elements of the imaging device 5, variations in intensity of illumination light, aberration of a lens, and the like. It may have happened. A in FIG. 11A is an example.

  In normal processing, the pixel value of the composite image 302 is binarized with a threshold, and the pixel whose pixel value is equal to or larger than the threshold is extracted as a defect candidate, and whether the defect candidate is a true defect due to the size of the area or the like In many cases, the determination is performed, but as described above, since the conventional method performs comparison processing with the threshold value with respect to the pixel values of all the pixels of the composite image 302, it takes time for calculation, and processing speed is slow. linked.

  In addition, the threshold for binarization needs to be set carefully in consideration of noise, and is easily influenced by the experience of the operator. In practice, in order to prevent erroneous extraction of a portion where the pixel value is large (without a true defect) due to noise as a defect candidate, a high value is often used as shown in B, and a true defect is a defect candidate When it is extracted as a factor, the size becomes smaller and the accuracy of defect detection decreases.

  Therefore, in the present embodiment, in S9, the pixel value of each pixel of the composite image 302 is uniformly subtracted from the pixel value A of the pixel, and the pixel value of the pixel after calculation becomes less than 0 as indicated by A ′. The value is 0. This predetermined value is previously determined so as to make the value after calculation in a portion where only noise occurs (without a true defect) be less than zero.

  The inspection apparatus 3 performs a detection process of the stain defect a on the composite image 302 after the noise removal process of S9 (S10). Here, as shown in FIG. 11B, the pixel value of the composite image 302 is compared with a predetermined threshold C to be binarized, and a pixel having a pixel value equal to or higher than the threshold C is extracted as a defect candidate 303. Then, the area of the defect candidate 303 is compared with a reference value to determine whether the defect candidate 303 is a true defect. For example, if the area of the defect candidate 303 is equal to or larger than the reference value, the defect candidate 303 is regarded as a true defect.

  In this case, since the threshold C can be set without considering the noise, it is not influenced by the experience of the operator etc. The threshold C is set to a low value to extract a large number of defect candidates and improve the accuracy of defect detection. It will be possible to Further, the portion where the pixel value A 'is 0 is excluded from the analysis target, and the comparison between the pixel value A' and the threshold C may be performed only for the portion where the pixel value A 'is not 0. In addition, since the calculation of subtracting the predetermined value uniformly from the pixel value of each pixel can be performed in S9 in a very short time, the entire calculation does not take time, and the processing speed can be improved.

  In the present embodiment, the contamination defect a is detected from the inspection image 200 by the above processing. The mask images 301a and 301b and the composite image 302 described above correspond to the “image created based on the difference image” in the present invention.

  Next, inspection of the missing defect b will be described. The inspection apparatus 3 first creates a difference image 400 obtained by subtracting the pixel value of the reference image 100 from the pixel value of the inspection image 200 as the inspection of the missing defect b (S11).

  FIG. 12 shows an example of the difference image 400. In S11, calculation is performed to subtract the pixel value of the pixel of the reference image 100 from the pixel value of the pixel of the inspection image 200 between the pixels at corresponding positions of the inspection image 200 and the reference image 100. If the value after calculation is less than 0, the pixel value of that pixel is set to 0. By subtracting the pixel value of the reference image 100 from the pixel value of the inspection image 200, the defect b appears in the difference image 400. Further, along with the positional deviation of both the images 100 and 200, a part of the edge portion of the pattern also appears in the difference image 400.

  The inspection apparatus 3 also performs mask processing on the difference image 400 using the edge expansion image 202 to create a mask image 401 a (first mask image) (S 12), and uses the edge expansion image 102. A mask process is performed to create a mask image 401 b (second mask image) (S 13). FIGS. 13A and 13B show examples of the mask images 401a and 401b after the mask processing with the edge expansion images 202 and 102 is performed on the difference image 400, respectively, and the above-described edge portions are removed. The details of the mask process are the same as those in S6 and S7.

  Then, the inspection apparatus 3 combines the mask images 401a and 401b to create a combined image 402 (S14). FIG. 14 shows an example of the composite image 402. The details of the synthesis process are also the same as in the above-described S8.

  The inspection apparatus 3 also performs noise removal by subtracting a predetermined value from the pixel value of each pixel and setting the pixel value of the pixel whose pixel value is less than 0 to 0 also in the composite image 402 (S15). The details of the noise removal process are also the same as those in S9.

  The inspection apparatus 3 performs the detection process of the missing defect b on the composite image 402 after the noise removal process of S15 (S16). The details of the detection process of the missing defect b are also the same as those in S10 described above.

  As described above, the missing defect b in the inspection image 200 can be extracted separately from the dirt defect a. The mask images 401a and 401b and the composite image 402 described above also correspond to the "image created based on the difference image" in the present invention.

  As described above, in the present embodiment, noise is removed by the above-described method when defects are detected from the composite images 302 and 402 in S10 and S16. There is no need. Further, the part where the pixel value is 0 at the time of noise removal does not have to be considered at the time of detection of a defect, and can be excluded from the analysis target, so that the processing speed can be increased. Therefore, the convenience of the inspection can be improved.

  At the time of detection of a defect, pixel values of the combined image 302 and 402 after noise removal can be compared with the threshold C to extract a defect candidate. At this time, as described above, noise may not be taken into consideration when setting the threshold C, and a portion with a pixel value of 0 can be excluded from the analysis target. There is no need to do it, and the processing speed can be realized.

  Further, in the present embodiment, mask processing is performed on the difference image 300 (400) using the edge expansion images 202 and 102 to create two mask images 301a and 301b (401a and 401b), and then these are combined. Then, since the composite image 302 (402) is created, even if an edge portion appears in the difference image 300 (400) due to positional deviation between the inspection image 200 and the reference image 100, this is erroneously detected as a defect. Can be prevented, and defects can be reliably detected by using the composite image 302 (402) obtained by combining the two mask images 301a and 301b (401a and 401b) for defect detection. In the present embodiment, performing the noise removal process on the composite image 302 (402) allows the noise removal process to be performed only once, thereby speeding up the process.

  However, the present invention is not limited thereto. For example, in the present embodiment, two differential images 300 and 400 are created to separately detect the dirt defect a and the drop-out defect b, but only one of them may be created.

  Further, although the noise removal is performed on the combined image 302 (402) in the present embodiment, the noise removal may be performed on the mask images 301a and 301b (401a and 401b) before combining. In this case, the composite image 302 (402) for detecting a defect corresponds to the "image created based on the image after noise removal" in the present invention.

  Furthermore, in some cases, noise removal and defect detection may be performed directly on either one of the mask images, in which case the creation of the other mask image and the process of combining the mask images become unnecessary. Furthermore, it is also possible to perform noise removal and defect detection directly on the difference image 300 (400) by omitting the mask processing itself.

  Moreover, the detection method of the defect in S10 or S16 is not specifically limited, either. In the present invention, no matter what detection method is used, it is not necessary to take noise into consideration when detecting a defect, and do not take into consideration, when detecting a defect, a portion where the pixel value is 0 when removing noise. Since it is good, the processing can be speeded up.

  The preferred embodiments of the present invention have been described above with reference to the accompanying drawings, but the present invention is not limited to such examples. It is apparent that those skilled in the art can conceive of various modifications or alterations within the scope of the technical idea disclosed in the present application, and of course these also fall within the technical scope of the present invention. It is understood.

1: Inspection system 3: Inspection device 5: Imaging device 10: Paper 11: Difference image creation unit 12: Mask processing unit 13: Image composition unit 14: Noise removal unit 15: Defect detection unit 100: Reference image 101, 201: Edge Image 102, 202: edge expansion image 200: inspection image 300, 400: difference image 301a, 301b, 401a, 401b: mask image 302, 402: composite image 303: defect candidate

Claims (5)

An inspection apparatus for detecting a defect of the inspection object by comparing an inspection image obtained by photographing the inspection object with a reference image,
A difference image generation unit that generates a difference image by calculating a difference between pixel values of the inspection image and the reference image;
A noise removing unit that subtracts noise by subtracting a predetermined value from the pixel value of the difference image or an image created based on the difference image, and setting the pixel value of the pixel whose pixel value is less than 0 to 0;
A defect detection unit that detects a defect using the image after the noise removal or the image created based on the image;
An inspection apparatus comprising: The defect detection unit
The pixel value of the image after the noise removal or the image created based on the image is binarized with a predetermined threshold to extract a defect candidate, and the defect candidate is true based on the area of the defect candidate The inspection apparatus according to claim 1, wherein it is determined whether or not there is a defect of An edge expansion image obtained by expanding an edge of the inspection image is used to perform a mask process for removing an edge portion generated in the difference image to create a first mask image, and expand an edge of the reference image. The image processing apparatus further includes a mask processing unit that performs a mask process of removing an edge portion generated in the difference image using an edge expansion image to create a second mask image,
The noise removing unit
3. The inspection apparatus according to claim 1, wherein noise removal is performed by subtracting the predetermined value from the pixel value of the combined image obtained by combining the first and second mask images. An inspection method for detecting a defect of the inspection object by comparing an inspection image obtained by photographing the inspection object with a reference image,
The computer is
Calculating a difference between pixel values of the inspection image and the reference image;
Performing noise removal by subtracting a predetermined value from the pixel value of the difference image or the image created based on the difference image, and setting the pixel value of the pixel whose pixel value is less than 0 to 0;
Detecting a defect using the image after the noise removal or the image created based on the image;
An inspection method characterized by performing. Computer,
An inspection apparatus for detecting a defect of the inspection object by comparing an inspection image obtained by photographing the inspection object with a reference image,
A difference image generation unit that generates a difference image by calculating a difference between pixel values of the inspection image and the reference image;
A noise removing unit that subtracts noise by subtracting a predetermined value from the pixel value of the difference image or an image created based on the difference image, and setting the pixel value of the pixel whose pixel value is less than 0 to 0;
A defect detection unit that detects a defect using the image after the noise removal or the image created based on the image;
A program for functioning as an inspection device having.