Circuit board inspection method, circuit board inspection apparatus, and circuit board inspection program

Abstract

A substrate inspection method according to the present invention involves using an inspection image obtained by capturing an image of a substrate to be inspected to inspect the substrate. The method involves: generating a plurality of component images in a color space from a plurality of normal images and a plurality of normal difference images, and generating respective preprocessed images for the plurality of normal images on the basis of distribution information about the distribution of respective component values for the pixels in the plurality of component images; generating an inspection difference image that is the difference between the inspection image and an average image, generating a plurality of component images in a color space from the inspection image and the inspection difference image, and generating a preprocessed image for the inspection image on the basis of distribution information about the distribution of respective component values for the pixels in the plurality of component images; and inspecting the substrate for defects on the basis of the preprocessed images for the normal images and the preprocessed image for the inspection image.

Description

【Technical Field】 【0001】 The present disclosure relates to a substrate inspection method, a substrate inspection apparatus, and a substrate inspection program. 【Background Art】 【0002】 Patent Document 1 discloses an apparatus that classifies defects occurring on a substrate based on a captured image, which is an inspection target obtained by capturing the substrate. 【Prior Art Documents】 【Patent Documents】 【0003】 【Patent Document 1】 Japanese Unexamined Patent Application Publication No. 2019 - 124591 【Summary of the Invention】 【Problems to be Solved by the Invention】 【0004】 The present disclosure provides a technique useful for accurately detecting abnormalities on the surface of a substrate. 【Means for Solving the Problems】 【0005】 A substrate inspection method according to one aspect of the present disclosure is a substrate inspection method that inspects a substrate to be inspected using an inspection image obtained by imaging the substrate to be inspected, and includes: generating an average image generated from a plurality of normal images obtained by imaging a plurality of normal substrates, and a plurality of normal difference images which are the difference between each of the plurality of normal images; generating a plurality of component images in a color space from the plurality of normal images and the plurality of normal difference images; generating a preprocessed image for each of the plurality of normal images based on variance information relating to the variance of the component values ​​of each pixel included in the plurality of component images; generating an inspection difference image which is the difference between the inspection image and the average image; generating a plurality of component images in a color space from the inspection image and the inspection difference image; generating a preprocessed image for the inspection image based on variance information relating to the variance of the component values ​​of each pixel included in the plurality of component images; and inspecting defects in the substrate to be inspected based on the preprocessed image for the normal image and the preprocessed image for the inspection image. [Effects of the Invention] 【0006】 This disclosure provides a technology useful for accurately detecting abnormalities on the substrate surface. [Brief explanation of the drawing] 【0007】 [Figure 1] Figure 1 is a schematic perspective view showing the substrate processing system according to the first embodiment. [Figure 2] Figure 2 is a schematic side view showing an example of a coating and developing apparatus. [Figure 3] Figure 3 is a schematic diagram showing an example of an inspection unit. [Figure 4] Figure 4 is a block diagram showing an example of the functional configuration of a control device. [Figure 5] Figure 5 is a block diagram showing an example of the hardware configuration of a control device. [Figure 6] Figure 6 is a flowchart showing an example of a substrate inspection method. [Figure 7]Figure 7 is a schematic diagram illustrating an example of how to create a model map. [Figure 8] Figure 8 is a schematic diagram illustrating an example of an image processing method for a normal image. [Figure 9] Figures 9(a), 9(b), and 9(c) are schematic diagrams illustrating an example of an image processing method for a target image. [Figure 10] Figures 10(a), 10(b), 10(c), and 10(d) are schematic diagrams illustrating an example of an image processing method for a target image. [Figure 11] Figure 11 is a schematic diagram illustrating an example of a method for creating pre-processed images from examination images. [Figure 12] Figure 12 is a schematic diagram illustrating an example of a defect detection method. [Modes for carrying out the invention] 【0008】 The following describes various exemplary forms. 【0009】 A substrate inspection method according to an exemplary embodiment of the present disclosure is a substrate inspection method that performs inspection of a substrate to be inspected using an inspection image obtained by imaging the substrate to be inspected, and includes: generating an average image generated from a plurality of normal images obtained by imaging a plurality of normal substrates, and a plurality of normal difference images which are the difference between each of the plurality of normal images; generating a plurality of component images in a color space from the plurality of normal images and the plurality of normal difference images; generating a preprocessed image for each of the plurality of normal images based on variance information relating to the variance of the component values ​​of each pixel included in the plurality of component images; generating an inspection difference image which is the difference between the inspection image and the average image; generating a plurality of component images in a color space from the inspection image and the inspection difference image; generating a preprocessed image for the inspection image based on variance information relating to the variance of the component values ​​of each pixel included in the plurality of component images; and performing an inspection of defects in the substrate to be inspected based on the preprocessed image for the normal image and the preprocessed image for the inspection image. 【0010】 According to the above substrate inspection method, preprocessed images are prepared for each of the multiple normal images based on dispersion information relating to the dispersion of component values ​​of each pixel in the multiple component images in the color space generated from the multiple normal images. Similarly, a preprocessed image is prepared for the inspection image based on dispersion information relating to the dispersion of component values ​​of each pixel in the multiple component images in the color space generated from the inspection image of the substrate to be inspected. Then, the substrate to be inspected is inspected for defects based on these images. In this way, preprocessed images are generated using dispersion information relating to the dispersion of component values ​​of multiple components in the color space of the normal and inspection images, and these are used to inspect the substrate for defects. This allows for inspection using images that take into account the dispersion of component values ​​in the color space. Therefore, it is useful for accurately detecting abnormalities on the substrate surface. 【0011】 In this case, when generating a preprocessed image relating to the normal image, a model map may be created based on the dispersion information of the component values ​​of each pixel included in the multiple component images relating to the multiple normal images, and a preprocessed image relating to one normal image may be created by calculating the difference between the dispersion information of the component values ​​of each pixel included in the multiple component images obtained from one normal image and the model map. When generating a preprocessed image relating to the inspection image, a preprocessed image relating to the inspection image may be created by calculating the difference between the dispersion information of the component values ​​of each pixel included in the multiple component images obtained from the inspection image and the model map. 【0012】 In the configuration described above, a single model map created based on the dispersion information of the component values ​​of each pixel contained in multiple component images related to multiple normal images can be said to be a compilation of the dispersion of the component values ​​of each pixel in multiple normal images. By using such a model map to create pre-processed images related to normal and inspection images, it is possible to create pre-processed images that take into account the dispersion of component values ​​that may occur in normal images, thereby enabling more accurate detection of abnormalities on the substrate surface. 【0013】 The dispersion information may be obtained by calculating the Mahalanobis distance related to the component values of each pixel included in the plurality of component images. 【0014】 With such a configuration, by calculating the Mahalanobis distance, the dispersion information can be obtained more accurately, so that a preprocessing image more useful for detecting abnormalities on the substrate surface can be created. 【0015】 As the color space, at least one of the RGB color space, HSV color space, XYZ color space, and Lab color space may be used. 【0016】 As described above, by creating component images using at least one of the RGB color space, HSV color space, XYZ color space, and Lab color space, a preprocessing image more useful for detecting abnormalities on the substrate surface can be created. 【0017】 The normal difference image and the inspection difference image may include images after performing enhancement processing for enhancing the difference from the average image. 【0018】 The normal difference image and the inspection difference image may include images after performing unevenness removal processing for removing unevenness in the image. 【0019】 The normal difference image and the inspection difference image may include images after performing a process of obtaining the difference from the averaged value after averaging the component values of pixels in a predetermined region. 【0020】 The normal difference image and the inspection difference image may include images after performing noise removal processing for removing noise in the image. 【0021】 The inspection of defects in the substrate to be inspected may include the following steps: creating a normal image abnormality map showing the degree of abnormality for each pixel in the preprocessed image relating to the normal image, and an inspection image abnormality map showing the degree of abnormality for each pixel in the preprocessed image relating to the inspection image; and identifying a region where defects are presumed to exist based on the difference between the normal image abnormality map and the inspection image abnormality map. 【0022】 With the above configuration, regions where defects are presumed to exist are identified from the difference between the normal image abnormality map created from the preprocessed image related to the normal image and the inspection image abnormality map created from the preprocessed image related to the inspection image. The abnormality map is a map that shows the abnormality of each pixel, and since regions where defects exist are estimated from the difference in the abnormality of each pixel between the normal image and the inspection image, estimation can be performed with greater accuracy based on the information of each pixel. 【0023】 In creating the normal image abnormality map and the inspection image abnormality map, an algorithm is used that generates an abnormality map by inputting a preprocessed image, and this algorithm may be created by machine learning using the plurality of normal images, the plurality of normal difference images, and at least a portion of the plurality of component images obtained therefrom as training data. 【0024】 With the above configuration, the algorithm for generating an anomaly map is created using machine learning with training data consisting of multiple normal images, multiple normal difference images, and at least a portion of the multiple component images obtained from these. This configuration reduces the number of normal images that need to be prepared for algorithm creation, making it possible to create a more accurate algorithm more easily. 【0025】 A substrate inspection apparatus according to an exemplary embodiment of the present disclosure is a substrate inspection apparatus that inspects a substrate to be inspected using an inspection image obtained by imaging the substrate to be inspected, and includes: a normal preprocessed image creation unit that generates an average image from a plurality of normal images obtained by imaging a plurality of normal substrates, a plurality of normal difference images which are the difference between each of the plurality of normal images, a plurality of component images in a color space from the plurality of normal images and the plurality of normal difference images, and generates a preprocessed image related to the normal image based on variance information relating to the variance of the component values ​​of each pixel included in the plurality of component images; an inspection preprocessed image creation unit that generates an inspection difference image which is the difference between the inspection image and the average image, a plurality of component images in a color space from the inspection image and the inspection difference image, and generates a preprocessed image related to the inspection image based on variance information relating to the variance of the component values ​​of each pixel included in the plurality of component images; and a defect detection unit that inspects defects in the substrate to be inspected based on the preprocessed image related to the normal image and the preprocessed image related to the inspection image. 【0026】 According to the above-described substrate inspection apparatus, preprocessed images are prepared for each of the multiple normal images based on dispersion information relating to the dispersion of component values ​​of each pixel in the multiple component images in the color space generated from the multiple normal images. Similarly, a preprocessed image is prepared for the inspection image based on dispersion information relating to the dispersion of component values ​​of each pixel in the multiple component images in the color space generated from the inspection image of the substrate to be inspected. Then, based on these images, the substrate to be inspected is inspected for defects. In this way, preprocessed images are generated using dispersion information relating to the dispersion of component values ​​of multiple components in the color space of the normal and inspection images, and these are used to inspect the substrate for defects. This allows for inspection using images that take into account the dispersion of component values ​​in the color space. Therefore, it is useful for accurately detecting abnormalities on the substrate surface. 【0027】 The normal preprocessing image creation unit may generate a model map based on the dispersion information of the component values ​​of each pixel included in the multiple component images relating to the multiple normal images, and create a preprocessed image relating to one of the normal images by calculating the difference between the dispersion information of the component values ​​of each pixel included in the multiple component images obtained from one of the normal images and the model map. The inspection preprocessing image creation unit may create a preprocessed image relating to the inspection image by calculating the difference between the dispersion information of the component values ​​of each pixel included in the multiple component images obtained from the inspection image and the model map. 【0028】 The aforementioned dispersion information may be obtained by calculating the Mahalanobis distance related to the component value of each pixel included in the plurality of component images. 【0029】 The defect detection unit may create a normal image abnormality map showing the degree of abnormality for each pixel in the preprocessed image relating to the normal image, and an inspection image abnormality map showing the degree of abnormality for each pixel in the preprocessed image relating to the inspection image, and identify a region where a defect is presumed to exist from the difference between the normal image abnormality map and the inspection image abnormality map. 【0030】 The defect detection unit uses an algorithm to generate an abnormality map by inputting a preprocessed image when creating the normal image abnormality map and the inspection image abnormality map. This algorithm may be created by machine learning using the multiple normal images, the multiple normal difference images, and at least a portion of the multiple component images obtained therefrom as training data. 【0031】 A substrate inspection program according to an exemplary form of the present disclosure is a substrate inspection program that causes a computer to perform a substrate inspection using an inspection image obtained by imaging the substrate to be inspected, and causes the computer to perform the following: generate an average image generated from a plurality of normal images obtained by imaging a plurality of normal substrates, and a normal difference image which is the difference between each of the plurality of normal images; generate a plurality of component images in a color space from the plurality of normal images and the plurality of normal difference images; generate a preprocessed image for each of the plurality of normal images based on variance information relating to the variance of the component values ​​of each pixel included in the plurality of component images; generate an inspection difference image which is the difference between the inspection image and the average image; generate a plurality of component images in a color space from the inspection image and the inspection difference image; generate a preprocessed image for the inspection image based on variance information relating to the variance of the component values ​​of each pixel included in the plurality of component images; and perform an inspection for defects in the substrate to be inspected based on the preprocessed image for the normal image and the preprocessed image for the inspection image. 【0032】 According to the above circuit board inspection program, the same effect as the above circuit board inspection method is achieved. 【0033】 [Exemplary Embodiments] Various exemplary embodiments will be described in detail below with reference to the drawings. The same or corresponding parts will be denoted by the same reference numerals in each drawing. Some drawings show a Cartesian coordinate system defined by the X, Y, and Z axes. In the following description, the Z axis corresponds to the vertical direction, and the X and Y axes correspond to the horizontal direction. 【0034】 [Circuit board processing system] The substrate processing system 1 shown in Figure 1 is a system that performs the following processes on a workpiece W: formation of a photosensitive film, exposure of the photosensitive film, and development of the photosensitive film. The workpiece W to be processed is, for example, a substrate, or a substrate in which a film or circuit has been formed by a predetermined process. One example of a substrate included in the workpiece W is a silicon wafer. The workpiece W (substrate) may be formed in a circular shape. The workpiece W to be processed may be a glass substrate, a mask substrate, or an FPD (Flat Panel Display), or an intermediate obtained by performing a predetermined process on such a substrate. The photosensitive film is, for example, a resist film. 【0035】 The substrate processing system 1 comprises a coating and developing apparatus 2 and an exposure apparatus 3. The exposure apparatus 3 performs exposure processing on a resist film (photosensitive coating) formed on a workpiece W (substrate). Specifically, the exposure apparatus 3 irradiates the portion of the resist film to be exposed with energy rays using methods such as immersion exposure. The coating and developing apparatus 2 performs a process to form a resist film on the surface of the workpiece W before the exposure processing by the exposure apparatus 3, and then performs development processing on the resist film after the exposure processing. 【0036】 (Substrate processing equipment) The following describes the configuration of a coating and developing apparatus 2 as an example of a substrate processing apparatus. As shown in Figures 1 and 2, the coating and developing apparatus 2 comprises a carrier block 4, a processing block 5, an interface block 6, and a control device 100. 【0037】 The carrier block 4 introduces the workpiece W into the coating and developing apparatus 2 and takes the workpiece W out of the coating and developing apparatus 2. For example, the carrier block 4 can support multiple carriers C (storage sections) for the workpiece W and incorporates a transport device A1 including a transfer arm. The carrier C accommodates, for example, multiple circular workpieces W. The transport device A1 takes the workpiece W from the carrier C and passes it to the processing block 5, and receives the workpiece W from the processing block 5 and returns it to the carrier C. The processing block 5 has multiple processing modules 11, 12, 13, and 14. 【0038】 The processing module 11 incorporates a liquid treatment unit U1, a heat treatment unit U2, an inspection unit U3, and a transport device A3 for transporting the workpiece W to these units. The processing module 11 forms an underlayer film on the surface of the workpiece W using the liquid treatment unit U1 and the heat treatment unit U2. The liquid treatment unit U1 of the processing module 11 applies a treatment liquid for underlayer film formation onto the workpiece W. The heat treatment unit U2 of the processing module 11 performs various heat treatments associated with the formation of the underlayer film. The inspection unit U3 performs a process to inspect the surface condition of the workpiece W before the underlayer film is formed, after the underlayer film is formed, or before the treatment liquid for underlayer film formation is applied and heat treatment is performed. 【0039】 The processing module 12 incorporates a liquid processing unit U1, a heat processing unit U2, an inspection unit U3, and a transport device A3 for transporting the workpiece W to these units. The processing module 12 forms a resist film on the underlying film using the liquid processing unit U1 and the heat processing unit U2. The liquid processing unit U1 of the processing module 12 applies a processing liquid (resist) for resist film formation onto the underlying film. The heat processing unit U2 of the processing module 12 performs various heat treatments associated with the formation of the resist film. The inspection unit U3 performs a process to inspect the surface condition of the workpiece W before the resist film is formed, after the resist film is formed, or before the resist is applied and heat treatment is performed. 【0040】 The processing module 13 incorporates a liquid processing unit U1, a heat processing unit U2, an inspection unit U3, and a transport device A3 for transporting the workpiece W to these units. The processing module 13 forms an upper layer film on the resist film using the liquid processing unit U1 and the heat processing unit U2. The liquid processing unit U1 of the processing module 13 applies a processing liquid for upper layer film formation onto the resist film. The heat processing unit U2 of the processing module 13 performs various heat treatments associated with the formation of the upper layer film. The inspection unit U3 performs a process to inspect the surface condition of the workpiece W before the formation of the upper layer film, after the formation of the upper layer film, or before the processing liquid for upper layer film formation is applied and heat treatment is performed. 【0041】 The processing module 14 incorporates a liquid processing unit U1, a heat processing unit U2, an inspection unit U3, and a transport device A3 for transporting the workpiece W to these units. The processing module 14 performs development processing of the resist film after exposure using the liquid processing unit U1 and the heat processing unit U2. The liquid processing unit U1 of the processing module 14 performs development processing of the resist film by, for example, supplying a developer solution onto the surface of the exposed workpiece W and then washing it off with a rinsing solution. 【0042】 The heat treatment unit U2 of the processing module 14 performs various heat treatments associated with the development process. Specific examples of heat treatments include pre-development heating (PEB: Post Exposure Bake) and post-development heating (PB: Post Bake). The inspection unit U3 performs a process to inspect the surface condition of the workpiece W before development and PEB are performed, after development and PB are performed, or before the developer solution is supplied and PB is performed. 【0043】 A shelf unit U10 is provided on the carrier block 4 side within the processing block 5. The shelf unit U10 is divided into multiple cells arranged vertically. A transport device A7, including a lifting arm, is provided near the shelf unit U10. The transport device A7 raises and lowers the workpiece W between the cells of the shelf unit U10. 【0044】 A shelf unit U11 is provided on the interface block 6 side within the processing block 5. The shelf unit U11 is divided into multiple cells arranged vertically. 【0045】 The interface block 6 handles the transfer of workpieces W to and from the exposure device 3. For example, the interface block 6 incorporates a transport device A8, which includes a transfer arm, and is connected to the exposure device 3. The transport device A8 transfers the workpieces W placed on the shelf unit U11 to the exposure device 3, and receives the workpieces W from the exposure device 3 and returns them to the shelf unit U11. 【0046】 The control device 100 controls each device included in the coating and developing apparatus 2 to perform the coating and developing process (substrate processing) in the following procedure, for example. First, the control device 100 controls the transport device A1 to transport the workpiece W in the carrier C to the shelf unit U10, and then controls the transport device A7 to place the workpiece W into a cell for the processing module 11. 【0047】 Next, the control device 100 controls the transport device A3 to transport the workpiece W from the shelf unit U10 to the liquid treatment unit U1 in the processing module 11. The control device 100 controls the liquid treatment unit U1 to form a film of processing liquid for lower layer film formation on the surface of the workpiece W. The control device 100 controls the heat treatment unit U2 to heat the workpiece W, which has the processing liquid film for lower layer film formation formed on it, to form the lower layer film. After that, the control device 100 controls the transport device A3 to return the workpiece W with the lower layer film formed on it to the shelf unit U10, and controls the transport device A7 to place the workpiece W in a cell for the processing module 12. The control device 100 may also control the inspection unit U3 to inspect the surface of the workpiece W at any point during processing in the processing module 11. 【0048】 Next, the control device 100 controls the transport device A3 to transport the workpiece W from the shelf unit U10 to the liquid treatment unit U1 in the processing module 12. The control device 100 controls the liquid treatment unit U1 to form a film of processing liquid for resist film formation on the surface of the workpiece W. The control device 100 controls the heat treatment unit U2 to heat the workpiece W, which has a film of processing liquid for resist film formation formed on it, to form a resist film. After that, the control device 100 controls the transport device A3 to return the workpiece W to the shelf unit U10 and controls the transport device A7 to place the workpiece W in a cell for the processing module 13. The control device 100 may also control the inspection unit U3 to inspect the surface of the workpiece W at any point during processing in the processing module 12. 【0049】 Next, the control device 100 controls the transport device A3 to transport the workpiece W from the shelf unit U10 to the liquid treatment unit U1 in the processing module 13. The control device 100 also controls the liquid treatment unit U1 to form a film of processing liquid for upper film formation on the resist film of the workpiece W. The control device 100 controls the heat treatment unit U2 to heat the workpiece W, which has the processing liquid film for upper film formation formed on it, to form the upper film. After that, the control device 100 controls the transport device A3 to transport the workpiece W to the shelf unit U11. The control device 100 may also control the inspection unit U3 to inspect the surface of the workpiece W at any point during processing in the processing module 13. 【0050】 Next, the control device 100 controls the transport device A8 to send the workpiece W from the shelf unit U11 to the exposure device 3. Subsequently, the control device 100 controls the transport device A8 to receive the exposed workpiece W from the exposure device 3 and place it in the cell for the processing module 14 in the shelf unit U11. 【0051】 Next, the control device 100 controls the transport device A3 to transport the workpiece W from the shelf unit U11 to each unit in the processing module 14, and controls the liquid processing unit U1 and the heat processing unit U2 to perform the resist film development process on the workpiece W. After that, the control device 100 controls the transport device A3 to return the workpiece W to the shelf unit U10, and controls the transport devices A7 and A1 to return the workpiece W to the carrier C. The control device 100 may also control the inspection unit U3 to inspect the surface of the workpiece W at some point during processing in the processing module 14. This completes the coating and development process for one workpiece W. The control device 100 controls each unit of the coating and development apparatus 2 to perform the coating and development process for each of the subsequent workpieces W in the same manner as described above. 【0052】 The specific configuration of the substrate processing apparatus is not limited to the configuration of the coating and developing apparatus 2 illustrated above. The substrate processing apparatus may be any apparatus that includes a unit for inspecting the surface of the workpiece W to be subjected to a predetermined process, and a control device for controlling this unit. 【0053】 (Inspection unit) Next, the inspection unit U3 included in processing modules 11 to 14 will be described. The inspection unit U3 has the function of capturing image data of the surface of the workpiece W (hereinafter referred to as "surface Wa"). The inspection unit U3 may also capture image data of the entire surface Wa of the workpiece W. As shown in Figure 3, the inspection unit U3 includes, for example, a housing 30, a holding unit 31, a linear drive unit 32, an imaging unit 33, and a light projection / reflection unit 34. 【0054】 The holding unit 31 holds the workpiece W horizontally with its surface Wa facing upward. The linear drive unit 32 includes a power source, such as an electric motor, and moves the holding unit 31 along a horizontal, straight path. The imaging unit 33 has a camera 35, such as a CCD camera. The camera 35 is located near one end of the inspection unit U3 in the direction of movement of the holding unit 31 and is directed toward the other end in that direction of movement. The light-emitting / reflecting unit 34 emits light into the imaging range and guides the reflected light from that imaging range to the camera 35. For example, the light-emitting / reflecting unit 34 has a half mirror 36 and a light source 37. The half mirror 36 is located higher than the holding unit 31 and in the middle of the movement range of the linear drive unit 32, and reflects light from below to the camera 35. The light source 37 is located above the half mirror 36 and irradiates illumination light downward through the half mirror 36. 【0055】 The inspection unit U3 operates as follows to acquire image data of the surface Wa of the workpiece W. First, the linear drive unit 32 moves the holding unit 31. This causes the workpiece W to pass under the half mirror 36. During this passage process, reflected light from each part of the surface Wa of the workpiece W is sequentially sent to the camera 35. The camera 35 forms an image of the reflected light from each part of the surface Wa of the workpiece W and acquires image data of the surface Wa of the workpiece W (the entire surface Wa). The image obtained by imaging the surface Wa of the workpiece W changes depending on the state of the surface Wa of the workpiece W. Therefore, the inspection unit U3 acquires an image of the surface Wa of the workpiece W (image data) as information indicating the state of the surface Wa of the workpiece W, and uses this to evaluate the state of the surface Wa, especially the presence or absence of defects. 【0056】 The image data acquired by camera 35 is sent to control device 100. The control device 100 can inspect the condition of the surface Wa of the workpiece W based on the image data of the surface Wa. For example, it can inspect for defects on the surface Wa of the workpiece W. In this disclosure, image data in which the pixel value of each pixel is defined may simply be referred to as "image". 【0057】 [Control device (circuit board inspection device)] As shown in Figure 4, the control device 100 has a functional configuration (hereinafter referred to as "functional module") consisting of a processing control unit 102 and an inspection control unit 110. The processing performed by the processing control unit 102 and the inspection control unit 110 corresponds to the processing performed by the control device 100. The processing control unit 102 controls the liquid processing unit U1 and the heat processing unit U2 to perform the liquid processing and heat processing in the coating and developing process described above on the workpiece W. 【0058】 The inspection control unit 110 (substrate inspection device) inspects the workpiece W based on image data obtained from the inspection unit U3 at any stage during the coating and developing process. The inspection of the workpiece W includes determining whether or not there are any abnormalities (defects) on the surface Wa of the workpiece W. Examples of defects on the surface Wa include scratches, foreign matter adhesion, uneven application of the processing solution, and areas where the processing solution has not been applied. 【0059】 Before performing the inspection, the inspection control unit 110 prepares normal substrate data to be used for the inspection using a normal workpiece W (normal substrate) in which no defects were found on the surface Wa. Based on the normal substrate data, the inspection control unit 110 performs the inspection of the workpiece W (substrate to be inspected). The normal workpiece W and the workpiece W to be inspected are of the same type. The normal workpiece W and the workpiece W to be inspected are subjected to the same coating and developing process, and the preparation of the normal substrate data and the inspection of the workpiece W are performed at the same timing in the coating and developing process (for example, after the resist coating and before the heat treatment). 【0060】 The inspection control unit 110 has, as functional modules, a model map creation unit 120, a preprocessing unit 130, a defect detection unit 112, and a result output unit 114. Furthermore, the model map creation unit 120 has an image acquisition unit 122, a component image generation unit 124, and a model map generation unit 126. The preprocessing unit 130 has an image acquisition unit 132, a component image generation unit 134, and a preprocessed image generation unit 136. The model map creation unit 120 and the preprocessing unit 130 of the inspection control unit 110 function as a normal preprocessed image creation unit that creates preprocessed images related to normal images. Furthermore, the model map creation unit 120 and the preprocessing unit 130 of the inspection control unit 110 also function as an inspection preprocessed image creation unit that creates preprocessed images related to inspection images. The processing performed by each functional module of the inspection control unit 110 corresponds to the processing performed by the inspection control unit 110 (control device 100). 【0061】 The model map creation unit 120 has the function of creating a model map for inspection from images of a normal workpiece W. Multiple images (for example, 3 to 10 images) of the surface Wa of a normal workpiece W are prepared as image data related to the normal workpiece W. 【0062】 The image acquisition unit 122 has the function of acquiring image data of the surface Wa of a normal workpiece W from the inspection unit U3. 【0063】 The component image generation unit 124 has the function of generating component images used to create a model map from the image acquired by the image acquisition unit 122. The procedure for generating component images will be described later. 【0064】 The model map generation unit 126 has the function of creating a model map using the component image generated by the component image generation unit 124. A model map is a map of the shape corresponding to the workpiece W, obtained by quantifying the degree of color variation on the surface Wa in the image data of a normal workpiece W for each pixel included in the image data. One model map is created from multiple image data related to a normal workpiece W. The detailed procedure will be described later, but the color variation for each pixel is quantified using covariance and Mahalanobis distance. The model map created by the model map creation unit 120 is used for preprocessing of image data used for defect detection. 【0065】 The preprocessing unit 130 has the function of creating preprocessed images for use in inspection from images of a normal workpiece W and images of the workpiece W to be inspected. 【0066】 The image acquisition unit 132 has the function of acquiring image data of the surface Wa of the workpiece W to be inspected from the inspection unit U3. Since image data of a normal workpiece W surface Wa has already been acquired by the image acquisition unit 122, this image data is used. 【0067】 The component image generation unit 134 has the function of generating component images used to create preprocessed images from images acquired by the image acquisition unit 132. The procedure for generating component images will be described later. 【0068】 The preprocessing image generation unit 136 has the function of generating a preprocessing image using the component image generated by the component image generation unit 134. The preprocessing image is an image in which the degree of color variation on the surface Wa of the image data of a normal workpiece W and the image data of the workpiece W to be inspected is quantified for each pixel in the image data. The detailed procedure will be described later, but in the generation of the preprocessing image, the color variation for each pixel is quantified using the covariance used to create the model map and the Mahalanobis distance. One preprocessing image is created for each image data to be processed. A preprocessing image is also created for the image data of a normal workpiece W. The preprocessing image created by the preprocessing unit 130 is used in defect detection. 【0069】 The defect detection unit 112 uses image data of a normal workpiece W and a pre-processed image created from image data of the workpiece W to be inspected to check for defects on the surface Wa of the workpiece W to be inspected. An example of the processing procedure in the defect detection unit 112 will be described later. 【0070】 The result output unit 114 has the function of outputting the detection results from the defect detection unit 112. For example, if the defect detection unit 112 determines that there is an abnormality on the surface Wa of the workpiece W, it may output an abnormality signal indicating that the workpiece W being inspected is abnormal. When the result output unit 114 outputs an abnormality signal, it may output the abnormality signal to the processing control unit 102, to a higher-level controller, or to an output device such as a monitor for notifying operators of the information. 【0071】 The control device 100 is comprised of one or more computers. The control device 100 has, for example, the circuit 150 shown in Figure 5. The circuit 150 has one or more processors 152, a memory 154, a storage 156, and an input / output port 158. The storage 156 has a storage medium that can be read by the computer, such as a hard disk. The storage medium stores a program (board inspection program) that causes the control device 100 to execute the board inspection method described later. The storage medium may be a removable medium such as a non-volatile semiconductor memory, a magnetic disk, or an optical disk. 【0072】 Memory 154 temporarily stores the program loaded from the storage medium of storage 156 and the calculation results by processor 152. Processor 152 works in cooperation with memory 154 to execute the above program, thereby configuring each of the functional modules described above. Input / output ports 158 input and output electrical signals to and from the liquid treatment unit U1, heat treatment unit U2, inspection unit U3, etc., according to commands from processor 152. 【0073】 The hardware configuration of the control device 100 is not necessarily limited to configuring each functional module by program. For example, each functional module of the control device 100 may be composed of a dedicated logic circuit or an ASIC (Application Specific Integrated Circuit) that integrates them. If the control device 100 is composed of multiple computers (multiple circuits), some of the above functional modules may be implemented by one computer (circuit), and the remaining parts of the above functional modules may be implemented by other computers (circuits). 【0074】 [Circuit board inspection method] As an example of a substrate inspection method, a series of processes performed by the control device 100 (inspection control unit 110) will be described. Figure 6 is a flowchart illustrating the series of processes performed by the control device 100. Figures 7 to 12 are diagrams illustrating the details of the processes performed at each step and an example of the image data generated at each step. 【0075】 The control device 100 executes steps S01 to S03, for example, as shown in Figure 6. In step S01, the image acquisition unit 122 of the model map creation unit 120 acquires image data (normal image) relating to a normal workpiece W from the inspection unit U3. Next, in step S02, the component image generation unit 124 of the model map creation unit 120 creates component images from the normal image acquired by the image acquisition unit 122. Furthermore, in step S03, the model map generation unit 126 of the model map creation unit 120 generates a model map from the component images generated by the component image generation unit 124. 【0076】 Figure 7 illustrates the series of steps S01 to S03 described above. In step S01, it is assumed that there are three normal images G1 acquired by the image acquisition unit 122 and used in the procedure described later. For example, 2048 x 2048 pixel image data is used as the normal image G1. Also, both the normal image G1 (and the inspection image) are color images. In the case of an RGB color image, the pixel value information for each pixel is included, for example, information relating to each of RGB. Thus, one normal image G1 contains information relating to the pixel values ​​of 2048 x 2048 pixels. 【0077】 Figure 7 shows an example in which the image generation process P1 is performed on three normal images G1. The image generation process P1 shown in Figure 7 corresponds to step S02 in Figure 6, that is, the process of creating component images. 【0078】 Figure 7 shows image generation processes P1, including enhancement, difference processing, component decomposition, and filtering. The procedures for creating these component images will be explained with reference to Figures 8 to 10. First, an average image G2 is created by averaging the pixel values ​​of each pixel in multiple normal images G1. Next, enhancement processing is performed on the average image G2 to create an average enhanced image G3. As an example, enhancement processing using a LUT (Lookup Table) based on a grayscale conversion function is performed on a color image. A grayscale conversion function is a function that maps pixel values ​​in the input image and the output image, and a function is set for each RGB color according to the feature area to be emphasized. A LUT is a table that shows the mapping of pixel values ​​before and after conversion. In other words, by setting a grayscale conversion function that focuses on the area to be emphasized in advance and preparing a LUT corresponding to this grayscale conversion function, the RGB values ​​of each pixel in the color image are converted based on the LUT. As a result, for example, an average enhanced image G3 is obtained in which a region of specific pixel values ​​is more emphasized. 【0079】 Furthermore, a de-uniformity process is performed on this average-enhanced image G3. De-uniformity refers to a process that removes, for example, concentric circles of unevenness. Specifically, in the average-enhanced image G3, concentric circles are set with the pixel that captured the center of the workpiece W as the center. Then, by averaging the RGB pixel values ​​of the pixels on the concentric circles, a filter image can be prepared in which the RGB pixel values ​​of the concentric circles are the same (average value). Then, by taking the difference between the pixel values ​​of each pixel in the filter image and the pixel values ​​of each pixel in the average-enhanced image G3, an average de-uniformity image G4 is obtained. By performing this process, concentric circles of unevenness caused by different distances from the center can be removed. The average-enhanced image G3 and the average de-uniformity image G4 are obtained by the above procedure. 【0080】 Next, the image to be used to generate the component image is designated as the target image I1. By performing the same process on the target image as for generating the average enhanced image G3 described above, the target enhanced image I2 is generated. Furthermore, by performing the same process on the target enhanced image I2 as for generating the average uniformity removal image G4 described above, the target uniformity removal image I3 is generated. As a result, as shown in Figure 9(a), the target enhanced image I2 and the target uniformity removal image I3 are generated. Note that the target image I1 corresponds to one of the normal images G1 used to generate the average image G2 described above. 【0081】 Furthermore, as shown in Figure 9(b), a difference-enhanced image I4 is obtained by taking the difference between the target-enhanced image I2 and the average-enhanced image G3. Then, as shown in Figure 9(c), a difference-removed image I5 is obtained by taking the difference between the target-removed-blend image I3 and the average-removed-blend image G4. The difference-enhanced image I4 and the difference-removed-blend image I5 are images from which the components of the average-enhanced image G3 and the average-removed-blend image G4 have been removed, respectively. Therefore, the process of obtaining the difference-enhanced image I4 and the difference-removed-blend image I5 corresponds to the process of removing the fluctuating components contained in the average-enhanced image G3 and the average-removed-blend image G4, which were generated from an image obtained by averaging three normal images G1. In other words, if the normal image G1 is the target image I1, then both the difference-enhanced image I4 and the difference-removed-blend image I5 correspond to normal difference images. 【0082】 Furthermore, for each of the target enhancement image I2, target unevenness removal image I3, difference enhancement image I4, and difference unevenness removal image I5 obtained in the previous process, averaging, difference processing, and noise reduction processing are performed, as shown in Figures 10(a) to 10(d). Averaging processing is a process in which the pixel value of one pixel is set to the average value of the pixel values ​​of the surrounding group of pixels. For example, this can be done by using a 101x101 pixel mask to set the pixel value of one pixel to the average value of the surrounding 101x101 pixels. Next, difference processing is a process in which the color changes in large areas not caused by defects are removed by taking the difference between the image data after averaging processing and the image before processing. Furthermore, noise reduction processing is a process in which noise that is not thought to be related to defects is removed. For example, this can be done by using a 15x15 pixel mask to remove components caused by noise. By applying these processes, four processed images I11, I12, I13, and I14 are created from the target enhancement image I2, target unevenness removal image I3, difference enhancement image I4, and difference unevenness removal image I5, as shown in Figures 10(a) to 10(d). 【0083】 The above procedure is performed on an RGB color image. Therefore, the target enhancement image I2, the target unevenness removal image I3, and the four processed images I11, I12, I13, and I14 each contain information on the pixel values ​​of the R, G, and B components, respectively. By extracting only each component and generating image data, for example, from the target enhancement image I2 related to the RGB color image, we can obtain the target enhancement image I2 for the R component, the target enhancement image I2 for the G component, and the target enhancement image I2 for the B component. Similarly, images related to the R, G, and B components can be generated for the other images I3, I11 to I14. 【0084】 Furthermore, color images can be defined using a different color space instead of representing them with R, G, and B color components. Examples include the HSV color space, XYZ color space, or Lab color space. By representing the color of a single pixel using components from different color spaces, the characteristics of that pixel's color can be captured from a different perspective. Also, a color defined in a specific color space can be represented using components from another color space using a predetermined conversion formula. 【0085】 Therefore, a single color image is represented using the three components in the HSV color space (H, S, V), the XYZ color space (X, Y, Z), and the Lab color space (aa, ba, La). Then, for each of these components, images related to the target enhancement image I2, the target unevenness removal image I3, and the four processed images I11, I12, I13, and I14 are created. As a result, six types of images (target enhancement image I2, target unevenness removal image I3, and four processed images I11, I12, I13, and I14) are obtained by pre-processing each of the 12 color components (R, G, B, H, S, V, X, Y, Z, aa, ba, La) from a single color image. In other words, 72 component images are obtained based on 12 components × 6 types of processed images for a single color image. In Figure 7, component images G10 are shown as component images for the three normal images G1. 【0086】 Figure 7 shows that when the image generation process P1 is performed on three normal images G1, component images G10 are obtained. Furthermore, Figure 7 shows that a model map G20 is created from the component images G10. It also shows that the Mahalanobis distance is calculated in the creation of the model map G20. The procedure for creating the model map G20, including the calculation of the Mahalanobis distance, corresponds to step S03 in Figure 6, i.e., the model map creation process. 【0087】 The 72 component images G10 related to a single normal image G1 are thought to each contain different features. In other words, the 72 component images G10 can be considered to contain 72 different features, one for each pixel. In this case, assuming that 72 features are obtained for each pixel in a 2048 × 2048 pixel array, the mean and covariance matrix of each pixel are calculated, and then the Mahalanobis distance is calculated for each pixel. The mean and covariance matrix are calculated using the average values ​​of three images. The Mahalanobis distance for pixel M can be calculated, for example, using the following equation (1). In equation (1), μ is the mean and Σ is the covariance matrix. 【0088】 【number】 【0089】 By calculating the Mahalanobis distance, it is possible to determine how far the data of a given pixel is from the data set, i.e., the degree of data anomaly. Since the 72 features obtained from a single normal image G1 are different for each pixel, the degree to which the data of a given pixel is anomaly can be calculated using the Mahalanobis distance within a 2048×2048 pixel set. 【0090】 As described above, the Mahalanobis distance of each pixel in a normal image G1 can be calculated using 72 component images G10 obtained from a single normal image G1. By performing this process for each of the three normal images G1, three Mahalanobis distances can be calculated for the same pixel location from the three normal images G1. For each pixel, the maximum value of these three Mahalanobis distances is adopted and placed to correspond to a 2048 x 2048 pixel arrangement. The model map G20 is created in this way. In other words, the model map G20 can be said to be a mapping of the degree of abnormality of the data of each pixel contained in the three normal images G1. 【0091】 In the control device 100, after creating the model map G20 using the procedure described above, steps S04 to S07 are executed. In step S04, a preprocessed image related to the normal image G1 is prepared using the model map G20. In steps S05 to S07, a preprocessed image related to the image to be inspected is prepared using the model map G20. 【0092】 More specifically, in step S05, the image acquisition unit 132 of the preprocessing unit 130 acquires the image to be inspected. In step S06, the component image generation unit 134 of the preprocessing unit 130 generates component images of the image to be inspected. Then, in step S07, the preprocessing image generation unit 136 of the preprocessing unit 130 creates a preprocessed image related to the normal image G1 using the component image and model map G20 created in step S06. On the other hand, in step S04, the preprocessing image generation unit 136 of the preprocessing unit 130 creates a preprocessed image related to the normal image G1 using the component image G10 and model map G20 created in step S02. 【0093】 The above procedure will be explained with reference to Figure 11. Figure 11 shows the procedure for creating a preprocessed image T30 from the inspection image T1, which is the image to be inspected. First, an image generation process P1 is performed on the inspection image T1. The image generation process P1 shown in Figure 11 is the same as the image generation process P1 shown in Figure 7. Furthermore, the image generation process P1 corresponds to step S06 in Figure 6, that is, the component image creation process. 【0094】 In Figure 11, as in Figure 7, the image generation process P1 includes enhancement processing, difference processing, component decomposition processing, and filtering processing. The procedure for creating these component images is the same as the procedure described in the above embodiment. However, the average enhanced image G3 and the average unevenness removal image G4 are created from the normal image G1. Based on these images, using the inspection image T1 as the target image I1, the target enhanced image I2 and the target unevenness removal image I3 are generated by the procedure shown in Figure 9(a). Next, as shown in Figure 9(b), the difference enhanced image I4 is generated by taking the difference between the target enhanced image I2 and the average enhanced image G3. Furthermore, as shown in Figure 9(c), the difference unevenness removal image I5 is generated by taking the difference between the target unevenness removal image I3 and the average unevenness removal image G4. Thus, the difference enhanced image I4 and the difference unevenness removal image I5 created when the inspection image T1 is the target image I1 both correspond to inspection difference images. 【0095】 Then, for each of the target enhancement image I2, target unevenness removal image I3, difference enhancement image I4, and difference unevenness removal image I5 obtained in the process described above, averaging, difference processing, and noise reduction processing are performed, as shown in Figures 10(a) to 10(d), to create four processed images I11, I12, I13, and I14. By performing this processing on the 12 components described above, six types of images (target enhancement image I2, target unevenness removal image I3, and four processed images I11, I12, I13, and I14) are generated from a single color image, each obtained by preprocessing the 12 color components (R, G, B, H, S, V, X, Y, Z, aa, ba, La). As a result, 72 component images T10 are obtained from a single inspection image T1, based on 12 components × 6 types of processed images, as shown in Figure 11. 【0096】 Subsequently, as shown in Figure 11, a map image T20 is created from the component image T10. The Mahalanobis distance is calculated during the creation of the map image T20. The above formula (1) is used to calculate the Mahalanobis distance when creating the map image T20. However, the mean and covariance matrix of each pixel in the normal image G1 (the average value of the three images) is used. In other words, the procedure differs from that when creating the model map G20 in that the mean and covariance matrix of the normal image G1 is used, rather than the mean and covariance matrix derived from the inspection image T1. 【0097】 By calculating the Mahalanobis distance, it is possible to determine how far the data of a given pixel is from the data set, i.e., the degree of data anomaly. Since the 72 features obtained from a single inspection image T1 are different for each pixel, the degree to which the data of a given pixel is anomaly can be calculated using the Mahalanobis distance within a 2048×2048 pixel set. 【0098】 As described above, the Mahalanobis distance of each pixel in the inspection image T1 can be calculated using 72 component images T10 obtained from a single inspection image T1. The Mahalanobis distances of each pixel are then arranged to correspond to a 2048 x 2048 pixel layout. The resulting map image T20 is created in this way. In other words, the map image T20, like the model map G20, maps the degree of abnormality of the data of each pixel contained in the inspection image T1. 【0099】 Then, as shown in Figure 11, the preprocessed image T30 of the inspection image T1 is created by calculating the difference between the map image T20 and the model map G20. By performing difference processing using the model map G20, information related to the degree of abnormality that is shown in the model map G20 created from the normal image G1 is removed. The image after performing the process of calculating the difference between the map image T20 and the model map G20 becomes the preprocessed image T30. Note that the creation of the map image T20 and the preprocessed image T30 after preparing the component image T10 corresponds to the generation of the preprocessed image in step S07. 【0100】 By performing the above series of processes, a pre-processed image T30 is created from a single inspection image T1. 【0101】 In step S04, which creates a preprocessed image corresponding to the normal image G1, the preprocessed image generation unit 136 of the preprocessing unit 130 uses the component image G10 and model map G20 created in step S02 to create a preprocessed image for each of the three normal images G1. At this time, the component image G10 related to the three normal images G1 is prepared in the procedure for creating the model map G20. Therefore, the steps up to the preparation of the component image G10 in the procedure shown in Figure 11 can be omitted. Accordingly, for the normal image G1, the creation of the map image (corresponding to the map image T20 in Figure 11) and the creation of the preprocessed image (corresponding to the preprocessed image T30 in Figure 11) after the preparation of the component image G10 corresponds to the generation of the preprocessed image in step S07. Furthermore, by executing step S07, a preprocessed image G30 (see Figure 12) for each of the three normal images G1 can be created. 【0102】 After creating a preprocessed image G30 corresponding to the normal image G1 and a preprocessed image T30 corresponding to the inspection image T1 using the procedure described above, the control device 100 executes steps S11 to S14. In steps S11 and S12, the control device 100 extracts features from the preprocessed image G30 of the normal image G1 and calculates the degree of abnormality (step S11) to create an abnormality map used for defect detection (step S12). In steps S13 and S14, the control device 100 extracts features and calculates the degree of abnormality (step S13) to create an abnormality map used for defect detection (step S14). Steps S11 to S14 are performed by the defect detection unit 112 of the control device 100. 【0103】 The feature extraction performed in steps S11 and S13 is the process of extracting features related to defects from the preprocessed images G30 and T30, and can be performed using, for example, the Vision Transformer (ViT), one of the image classification algorithms. The anomaly score calculation performed in steps S12 and S14 is the process of calculating the anomaly score related to defects for each pixel based on the features from the preprocessed images G30 and T30, and can be performed using, for example, the Graph Mixture Density Networks (GMDN), one of the methods for determining probability distributions using neural networks. The same algorithm is used in both steps S11 and S13, and steps S12 and S14. 【0104】 Thus, algorithms generated by machine learning using training data can be used for feature extraction and anomaly score calculation. During the training phase of the algorithms used for feature extraction and anomaly score calculation, the algorithms are adjusted so that when features are extracted and anomaly scores are calculated from pre-processed images created from multiple normal images prepared in advance, the calculated anomaly score is reduced. By applying the algorithm (model) whose parameters have been adjusted through such pre-processing to pre-processed image G30 created from normal image G1 and pre-processed image T30 created from inspection image T1, the anomaly score for each pixel can be calculated. 【0105】 The defect detection unit 112 of the control device 100 maps the pixel-specific abnormality information calculated in the above procedure to the pixel arrangement. As a result, one abnormality map G40 (normal image abnormality map) is created from three normal images G1, and one abnormality map T40 (inspection image abnormality map) is created from one inspection image T1. 【0106】 Next, the control device 100 executes step S15. In step S15, the defect detection unit 112 of the control device 100 uses the abnormality maps G40 and T40 described above to determine whether or not there is a defect in the inspection image T1. The result output unit 114 of the control device 100 also outputs the result of the defect determination. 【0107】 Figure 12 shows an example of a procedure for determining the presence or absence of defects. First, as described above, a difference image T41 is obtained by calculating the difference between an abnormality map G40 created from three normal images G1 and an abnormality map T40 created from one inspection image T1. Furthermore, a binarized image T42 is prepared by binarizing the value of each pixel of this difference image T41 using a predetermined threshold, and a binarized image T43 is prepared by binarizing the value of each pixel of the preprocessed image T30 of the inspection image T1 created in the above procedure using a predetermined threshold. Then, the two binarized images T42 and T43 are superimposed to create a defect detection image T45 in which pixels that have been converted to white in both images are identified. In the defect detection image T45, the parts displayed in white are determined to have defects. Note that the above procedure is just one example, and the defect detection method is not particularly limited. 【0108】 By performing the above processing, regions where defects are presumed to exist are identified based on the inspection image T1. The result output unit 114 of the control device 100 may output the defect detection image T45 as the determination result, or it may output only the determination result of whether or not there is a defect based on whether or not the region determined to have a defect is included in the defect detection image T45. 【0109】 The above procedure uses a machine learning algorithm, but the learning method can be modified as appropriate. For example, as mentioned above, during the learning phase of the algorithm used for feature extraction and anomaly score calculation, adjustments are made so that the calculated anomaly score is low when features are extracted from preprocessed images created from multiple pre-prepared normal images. In the above example, preprocessed images created from normal images are used for learning, but instead of using only preprocessed images for learning, the component images used to create the preprocessed images may also be used as training data for machine learning. As mentioned above, preprocessed images generated from normal images are created using 72 component images created from one normal image. Since these component images contain information related to the feature parts of the normal image, they can be used as training data for the feature extraction and anomaly score calculation algorithms. Therefore, the feature extraction and anomaly score calculation algorithms may be performed with at least some of the component images used to generate the preprocessed images included in the training data. With this configuration, not only is one preprocessed image obtained from one normal image used as training data, but component images are also used. This allows for an increase in the amount of training data while reducing the number of normal images prepared in advance. Therefore, it becomes easier to prepare sufficient training data for effective learning. 【0110】 [effect] According to the above-described substrate inspection apparatus (inspection control unit 110) and substrate inspection method, a preprocessed image G30 for each of the multiple normal images G1 is prepared based on dispersion information relating to the dispersion of component values ​​of each pixel included in the multiple component images in the color space generated from the multiple normal images. Similarly, a preprocessed image T30 for the inspection image is prepared based on dispersion information relating to the dispersion of component values ​​of each pixel included in the multiple component images in the color space generated from the inspection image T1 of the substrate to be inspected. Then, based on these images, the substrate to be inspected is inspected for defects. In this way, preprocessed images G30 and T30 are generated using dispersion information relating to the dispersion of component values ​​of multiple components in the color space of the normal image G1 and the inspection image T1, and these are used to inspect the substrate for defects. This makes it possible to perform inspection using images that take into account the dispersion of component values ​​in the color space. Therefore, it is useful for accurately detecting abnormalities on the substrate surface. 【0111】 It has been known for some time that component images of the color space are used as a method for inspecting defects in substrates using images. However, in conventional methods, information including differences in color tone that may be present in each image is used for defect inspection, so there was room for improvement in terms of inspection accuracy. In contrast, according to the method described in the above embodiment, pre-processed images G30 and T30 used for inspection are created using information related to the variance of component values ​​obtained from the component image. With this configuration, pre-processed images G30 and T30 that reflect information about the variance of component values ​​contained in the component image can be obtained. By using such pre-processed images for defect inspection, the accuracy of defect detection can be improved. 【0112】 Furthermore, in the above embodiment, a single model map G20 is created based on the dispersion information of the component values ​​of each pixel included in multiple component images. Then, a preprocessed image G30 related to a single normal image is created by calculating the difference between the dispersion information of the component values ​​of each pixel included in multiple component images obtained from a single normal image and the model map G20. Then, a preprocessed image T30 related to an inspection image is created by calculating the difference between the dispersion information of the component values ​​of each pixel included in multiple component images obtained from an inspection image and the model map G20. A single model map can be said to be a compilation of the dispersion of the component values ​​of each pixel in multiple normal images G1. By using such a model map to create preprocessed images related to the normal image G1 and the inspection image T1, it is possible to create preprocessed images that take into account the dispersion of component values ​​that may occur in the normal image. Therefore, abnormalities on the substrate surface can be detected with greater accuracy. 【0113】 Furthermore, dispersion information may be obtained by calculating the Mahalanobis distance related to the component values ​​of each pixel included in multiple component images. With this configuration, dispersion information can be obtained with greater accuracy by calculating the Mahalanobis distance, making it possible to create a pre-processed image that is more useful when detecting anomalies on the substrate surface. 【0114】 As the color space, at least one of the RGB color space, HSV color space, XYZ color space, and Lab color space may be used. As described above, by creating a component image using at least one of the RGB color space, HSV color space, XYZ color space, and Lab color space, a pre-processed image that is more useful when detecting abnormalities on the substrate surface can be created. In addition, as described in the above embodiment, component images in multiple color spaces from the RGB color space, HSV color space, XYZ color space, and Lab color space may be used. In this case, a component image is obtained when one image is defined in a different color space. By creating a pre-processed image using such a component image, the information of one image is decomposed into multiple components that are different from each other, and a component image obtained is used, which allows for a more multifaceted capture of the features contained in the image. Therefore, a pre-processed image useful for improving the accuracy of defect detection can be created. 【0115】 The normal difference image and the inspection difference image may include images after enhancement processing to highlight the difference from the average image. This configuration allows for obtaining images that further emphasize the characteristic components of each image, which can then be used to create pre-processed images. 【0116】 The normal difference image and the inspection difference image may include images after an image removal process has been performed to remove irregularities in the image. With this configuration, it is possible to obtain an image from which irregularities occurring in the image have been appropriately removed, and this can be used to create a pre-processed image. 【0117】 The normal difference image and the inspection difference image may include an image obtained after averaging the component values ​​of pixels in a predetermined region and then calculating the difference from the averaged value. With this configuration, it is possible to obtain an image from which abnormal values ​​that occur incidentally in single pixels, etc., have been removed, and this can be used to create a pre-processed image. 【0118】 The normal difference image and the inspection difference image may include images after noise reduction processing has been performed to remove noise from the image. With this configuration, it is possible to obtain an image from which potential noise in the image has been removed, and this can be used to create a pre-processed image. 【0119】 Inspecting defects in a substrate may include creating a normal image abnormality map G40 that shows the degree of abnormality for each pixel in a preprocessed image G30 related to a normal image G1, and an inspection image abnormality map T40 that shows the degree of abnormality for each pixel in a preprocessed image T30 related to an inspection image T1, and identifying areas where defects are presumed to exist from the difference between the normal image abnormality map G40 and the inspection image abnormality map T40. With the above configuration, areas where defects are presumed to exist are identified from the difference between the normal image abnormality map G40 created from the preprocessed image G30 related to the normal image G1 and the inspection image abnormality map T40 created from the preprocessed image T30 related to the inspection image T1. Since the abnormality maps G40 and T40 are maps that show the degree of abnormality for each pixel, and the areas where defects exist are estimated from the difference in the degree of abnormality for each pixel between the normal image G1 and the inspection image T1, the estimation is performed with greater accuracy based on the information of each pixel. 【0120】 In creating the normal image abnormality map G40 and the examination image abnormality map T40, an algorithm that generates the abnormality map by inputting a preprocessed image may be used. In this case, the algorithm may be created by machine learning using multiple normal images G1, multiple normal difference images, and at least a portion of the multiple component images obtained therefrom as training data. With the above configuration, the algorithm for generating the abnormality map is created by machine learning using multiple normal images, multiple normal difference images, and at least a portion of the multiple component images obtained therefrom as training data. By adopting such a configuration, the number of normal images prepared for creating the algorithm can be reduced, making it possible to create a more accurate algorithm more easily. 【0121】 [Differentiation] Although various exemplary embodiments have been described above, the invention is not limited to the exemplary embodiments described above, and various omissions, substitutions, and modifications may be made. Furthermore, it is possible to combine elements from different embodiments to form other embodiments. 【0122】 For example, the above embodiment describes a case in which six types of images (target enhancement image I2, target unevenness removal image I3, and four processed images I11, I12, I13, I14) are created by preprocessing each of the 12 color components (R, G, B, H, S, V, X, Y, Z, aa, ba, La) from a single color image using the RGB color space, HSV color space, XYZ color space, and Lab color space. However, this configuration can be changed as appropriate, and for example, only a part of the above-mentioned color spaces may be used. Alternatively, component images may be created using a color space different from the above-mentioned color spaces. 【0123】 Furthermore, the procedure for creating an anomaly severity map and detecting defects after creating the preprocessed image can be modified as appropriate. The procedure described in the above embodiment is just one example and can be changed to various methods that can be used in defect detection using images. In addition, depending on the method that can be used in defect detection, further image processing may be applied to the preprocessed image. 【0124】 Furthermore, the series of processes described above can be modified as appropriate. In the series of processes described above, the control device 100 (inspection control unit 110) may execute one step and the next step in parallel, or it may execute each step in a different order than the example described above. The control device 100 (inspection control unit 110) may omit any of the steps, or it may execute a different process in any of the steps than the example described above. 【0125】 The computer constituting the inspection control unit 110 may be located elsewhere than the coating and developing apparatus 2. In this case, the control device 100 and the inspection control unit 110 may be connected to each other via wired or wireless communication. The control device 100 may acquire an image from the inspection unit U3 and then transmit the image to the inspection control unit 110. The inspection control unit 110 may also transmit information indicating the determination result of whether or not an abnormality is present to the control device 100. 【0126】 From the above description, it will be understood that the various embodiments of this disclosure are described herein for illustrative purposes and can be modified in various ways without departing from the scope and spirit of this disclosure. Accordingly, the various embodiments disclosed herein are not intended to be limiting, and the true scope and spirit are shown by the appended claims. [Explanation of Symbols] 【0127】 1...Substrate processing system, 2...Coating and developing apparatus, 100...Control device, 102...Processing control unit, 110...Inspection control unit, 112...Defect detection unit, 114...Result output unit, 120...Model map creation unit, 122...Image acquisition unit, 124...Component image generation unit, 126...Model map generation unit, 130...Preprocessing unit, 132...Image acquisition unit, 134...Component image generation unit, 136...Preprocessed image generation unit.

Claims

[Claim 1] A substrate inspection method that involves inspecting a substrate to be inspected using an inspection image obtained by imaging the substrate to be inspected, The process involves generating an average image from multiple normal images obtained by imaging multiple normal substrates, generating multiple normal difference images which are the differences between each of the multiple normal images, generating multiple component images in color space from the multiple normal images and the multiple normal difference images, and generating a preprocessed image for each of the multiple normal images based on the variance information relating to the variance of the component values ​​of each pixel included in the multiple component images. The process involves generating an inspection difference image, which is the difference between the inspection image and the average image; generating multiple component images in a color space from the inspection image and the inspection difference image; and generating a preprocessed image related to the inspection image based on the variance information relating to the variance of the component values ​​of each pixel included in the multiple component images. Based on the pre-processed image relating to the normal image and the pre-processed image relating to the inspection image, the inspection of defects in the substrate to be inspected is performed. A circuit board inspection method, including the following. [Claim 2] In generating a preprocessed image relating to the normal image, a model map is created based on the dispersion information of the component values ​​of each pixel included in the multiple component images relating to the multiple normal images, and a preprocessed image relating to one of the normal images is created by calculating the difference between the dispersion information of the component values ​​of each pixel included in the multiple component images obtained from one of the normal images and the model map. The substrate inspection method according to claim 1, wherein, in generating a preprocessed image relating to the inspection image, the preprocessed image relating to the inspection image is created by calculating the difference between the variance information of the component values ​​of each pixel included in the plurality of component images obtained from the inspection image and the model map. [Claim 3] The substrate inspection method according to claim 1, wherein the dispersion information is obtained by determining the Mahalanobis distance relating to the component value of each pixel included in the plurality of component images. [Claim 4] The substrate inspection method according to claim 1, wherein at least one of the RGB color space, HSV color space, XYZ color space, and Lab color space is used as the color space. [Claim 5] The substrate inspection method according to claim 1, wherein the normal difference image and the inspection difference image include images after enhancement processing has been performed to enhance the difference with the average image. [Claim 6] The substrate inspection method according to claim 1, wherein the normal difference image and the inspection difference image include images after an unevenness removal process has been performed to remove unevenness in the image. [Claim 7] The substrate inspection method according to claim 1, wherein the normal difference image and the inspection difference image include images obtained after performing a process to calculate the difference from the averaged values ​​after averaging the component values ​​of pixels in a predetermined region. [Claim 8] The substrate inspection method according to claim 1, wherein the normal difference image and the inspection difference image include images after noise reduction processing has been performed to remove noise in the images. [Claim 9] Performing the inspection of defects in the substrate subject to inspection is, To create a normal image abnormality map showing the degree of abnormality for each pixel in the preprocessed image relating to the normal image, and an inspection image abnormality map showing the degree of abnormality for each pixel in the preprocessed image relating to the inspection image, The difference between the normal image abnormality map and the inspection image abnormality map is used to identify areas where defects are presumed to exist. A substrate inspection method according to claim 1, including the method described in claim 1. [Claim 10] In creating the aforementioned normal image abnormality map and the aforementioned examination image abnormality map, an algorithm is used that generates an abnormality map by inputting a pre-processed image. The substrate inspection method according to claim 9, wherein the algorithm is created by machine learning using at least a portion of the plurality of normal images, the plurality of normal difference images, and the plurality of component images obtained therefrom as training data. [Claim 11] A substrate inspection apparatus that performs inspection of a substrate to be inspected using an inspection image obtained by imaging the substrate to be inspected, A normal preprocessing image creation unit generates an average image from multiple normal images obtained by imaging multiple normal substrates, generates multiple normal difference images which are the differences between each of the multiple normal images, generates multiple component images in color space from the multiple normal images and the multiple normal difference images, and generates a preprocessed image related to the normal image based on variance information relating to the variance of the component values ​​of each pixel included in the multiple component images, An inspection preprocessing image creation unit generates an inspection difference image which is the difference between the inspection image and the average image, generates multiple component images in color space from the inspection image and the inspection difference image, and generates a preprocessed image related to the inspection image based on variance information relating to the variance of the component values ​​of each pixel included in the multiple component images, A defect detection unit that performs inspection of defects in the substrate to be inspected based on the preprocessed image relating to the normal image and the preprocessed image relating to the inspection image, A circuit board inspection device, including a substrate inspection device. [Claim 12] The normal preprocessing image creation unit generates a model map based on the dispersion information of the component values ​​of each pixel included in the multiple component images relating to the multiple normal images, and creates a preprocessed image relating to one of the normal images by calculating the difference between the dispersion information of the component values ​​of each pixel included in the multiple component images obtained from one of the normal images and the model map. The substrate inspection apparatus according to claim 11, wherein the pre-inspection image creation unit creates a pre-processed image relating to the inspection image by calculating the difference between the variance information of the component values ​​of each pixel included in the plurality of component images obtained from the inspection image and the model map. [Claim 13] The substrate inspection apparatus according to claim 11, wherein the dispersion information is obtained by determining the Mahalanobis distance relating to the component value of each pixel included in the plurality of component images. [Claim 14] The defect detection unit is A normal image abnormality map showing the degree of abnormality for each pixel in the preprocessed image relating to the normal image and an inspection image abnormality map showing the degree of abnormality for each pixel in the preprocessed image relating to the inspection image are created. A substrate inspection apparatus according to claim 11, which identifies a region where a defect is presumed to exist based on the difference between the normal image abnormality map and the inspection image abnormality map. [Claim 15] The defect detection unit uses an algorithm that generates an abnormality map by inputting a preprocessed image when creating the normal image abnormality map and the inspection image abnormality map. The substrate inspection apparatus according to claim 14, wherein the algorithm is created by machine learning using at least a portion of the plurality of normal images, the plurality of normal difference images, and the plurality of component images obtained therefrom as training data. [Claim 16] A circuit board inspection program that causes a computer to perform a circuit board inspection using an inspection image obtained by imaging the circuit board to be inspected, The process involves generating an average image from multiple normal images obtained by imaging multiple normal substrates, generating multiple normal difference images which are the differences between each of the multiple normal images, generating multiple component images in color space from the multiple normal images and the multiple normal difference images, and generating a preprocessed image for each of the multiple normal images based on the variance information relating to the variance of the component values ​​of each pixel included in the multiple component images. The process involves generating an inspection difference image, which is the difference between the inspection image and the average image; generating multiple component images in a color space from the inspection image and the inspection difference image; and generating a preprocessed image related to the inspection image based on the variance information relating to the variance of the component values ​​of each pixel included in the multiple component images. Based on the pre-processed image relating to the normal image and the pre-processed image relating to the inspection image, the inspection of defects in the substrate to be inspected is performed. A circuit board inspection program that is executed by a computer.