Image inspection device, image inspection method, and image inspection program

Abstract

Provided is an image inspection device, an image inspection method, and an image inspection program that provide high-precision pass / fail judgment. [Solution] The system includes a learning dataset acquisition unit that acquires a learning dataset, a learning feature derivation unit that derives learning features of the learning image data, a clustering processing unit that clusters the learning image data based on the learning features to divide it into multiple clusters, an index derivation unit that derives, for each cluster, the center of gravity and distance of the multiple learning features belonging to the cluster and derives the average μ and variance σ of the multiple distances corresponding to the multiple learning features, an imaging feature derivation unit that extracts imaging features based on imaging data, a minimum distance derivation unit that derives each distance between the imaging features and the multiple centers of gravity and derives a minimum distance X that is the smallest of the distances, an anomaly calculation unit that calculates an anomaly degree α based on the minimum distance X, the average μ, and the variance σ, and a comparison unit that compares the anomaly degree α with a preset threshold.

Description

[Technical field]

[0001] The present invention relates to an image inspection device, an image inspection method, and an image inspection program. [Background technology]

[0002] Conventionally, in image inspection of products, in order to prevent defective products from being erroneously judged as good products, there has been an image inspection method that combines multiple judgment algorithms to make a final pass / fail judgment on the product (for example, Patent Document 1).

[0003] However, in conventional image inspection methods, the judgment scores were sometimes unstable, and it was not possible to clearly distinguish between good and bad products using a specific threshold value. Therefore, there was room for further improvement in the accuracy of the final pass / fail judgment of products. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent No. 6999150 Summary of the Invention [Problem to be solved by the invention]

[0005] The present invention aims to provide an image inspection device, an image inspection method, and an image inspection program that have high accuracy in determining pass / fail. [Means for solving the problem]

[0006] (1) An image inspection device according to one aspect of the present invention includes a training dataset acquisition unit that acquires a training dataset including a plurality of training image data; a training feature derivation unit that derives training features of the training image data; a clustering processing unit that clusters the training image data into a plurality of clusters based on the training features; an index derivation unit that derives, for each cluster, a center of gravity of the plurality of training features belonging to the cluster and a distance between the center of gravity and the training features, and derives an average μ and a variance σ of the plurality of distances corresponding to the plurality of training features; an imaging feature derivation unit that extracts imaging features based on imaging data captured by a camera; a minimum distance derivation unit that derives each distance between the imaging feature and the plurality of centers of gravity, and derives a minimum distance X that is the smallest of the distances; an anomaly calculation unit that calculates an anomaly degree α based on the minimum distance X, the average μ, and the variance σ; and a comparison unit that compares the anomaly degree α with a preset threshold. (2) In the above (1), an error processing unit may be provided that executes an error process when the abnormality degree α exceeds the threshold value. (3) In the above (1) or (2), the clustering processing unit may perform the clustering a plurality of times using random numbers, and the index derivation unit may derive the center of gravity for each cluster for each clustering, and may determine a representative cluster having a minimum residual sum of squares of the distances in the clusters as a population for calculating the minimum distance X. (4) In the above (1) or (2), the feature amount deriving unit may perform resizing processing, window function processing, FFT processing, or band-pass filter processing. (5) An image inspection method according to one aspect of the present invention includes a training dataset acquisition step of acquiring a training dataset including a plurality of training image data; a training feature derivation step of deriving training features of the training image data; a clustering processing step of clustering the training image data into a plurality of clusters based on the training features; an index derivation step of deriving, for each cluster, a center of gravity of the plurality of training features belonging to the cluster and a distance between the center of gravity and the training features, and deriving an average μ and a variance σ of the plurality of distances corresponding to the plurality of training features; an imaging feature derivation step of extracting imaging features based on imaging data captured by a camera; a minimum distance derivation step of deriving each distance between the imaging feature and the plurality of centers of gravity, and deriving a minimum distance X that is the smallest of the distances; an abnormality calculation step of calculating an abnormality degree α based on the minimum distance X, the average μ, and the variance σ; and a comparison step of comparing the abnormality degree α with a preset threshold. (6) An image inspection program according to one embodiment of the present invention causes a computer to execute a learning dataset acquisition function for acquiring a learning dataset including a plurality of learning image data; a learning feature derivation function for deriving learning features of the learning image data; a clustering processing function for clustering the learning image data into a plurality of clusters based on the learning features; an index derivation function for deriving, for each cluster, a center of gravity of the plurality of learning features belonging to the cluster and a distance between the center of gravity and the learning features, and deriving an average μ and a variance σ of the plurality of distances corresponding to the plurality of learning features; an imaging feature derivation function for extracting imaging features based on imaging data captured by a camera; a minimum distance derivation function for deriving each distance between the imaging feature and the plurality of centers of gravity, and deriving a minimum distance X that is the smallest of the distances; an abnormality calculation function for calculating an abnormality degree α based on the minimum distance X, the average μ, and the variance σ; and a comparison function for comparing the abnormality degree α with a preset threshold. Effect of the Invention

[0007] According to the present invention, it is possible to provide an image inspection device, an image inspection method, and an image inspection program that have high pass / fail judgment accuracy. [Brief description of the drawings]

[0008] [Figure 1] FIG. 1 is a diagram illustrating an overview of an image inspection device. [Diagram 2] FIG. 1 is a diagram for explaining an outline of an image inspection method. [Diagram 3] FIG. 1 is a diagram illustrating a flow of an image inspection method. [Figure 4] FIG. 11 is a diagram illustrating a flow of a feature amount derivation step. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] (Embodiment) Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings. Fig. 1 is a diagram for explaining an overview of an image inspection device 100. Fig. 2 is a diagram for explaining an overview of an image inspection method. Note that, hereinafter, parts having common functions may be given the same reference numerals or symbols.

[0010] (Image inspection device) The image inspection device 100 according to the embodiment is used, for example, to inspect the appearance of products such as electronic components.

[0011] 1, the image inspection device 100 includes a control means S. The image inspection device 100 may include a camera C.

[0012] The camera C appropriately captures an image of a product (part) to be inspected. The captured image data (image data) is sent to the control means S. Note that image data captured by an external camera not included in the image inspection device 100 may be input to the control means S of the image inspection device 100, so that the image inspection device 100 can then perform inspection on the captured image data.

[0013] The control means S is a computer system including a CPU, a memory M (storage medium) such as a RAM, a ROM, and an auxiliary storage device, an external connection interface IF with an input / output device, and a bus B. The control means S may include a network interface. The memory M stores an image inspection program for causing the control means S (computer system) to execute the image inspection method according to this embodiment.

[0014] The CPU is an arithmetic circuit that controls the control means S. The CPU reads out a program stored in the ROM or auxiliary storage device to the RAM. The CPU executes various processes in the program read out to the RAM. The ROM stores system programs and the like used to control the control means S. The auxiliary storage device stores application programs and the like that execute various processes. The auxiliary storage device is, for example, an HDD or SSD. The external connection interface IF is an interface for connecting various devices to the control means S. The external connection interface IF connects, for example, a camera C, a display D, a keyboard K, and the like to the control means S. The network interface functions to communicate via a network based on the control of the CPU. The bus B connects the above-mentioned functional units that constitute the control means S so that they can communicate with each other.

[0015] The image inspection device 100 captures an image of the product to be inspected using a camera C, and based on the captured image data, a control means S determines whether the product to be inspected is pass / fail (whether the product is good or defective, or normal or abnormal).

[0016] As shown in FIG. 2, the image inspection device 100 performs inspection based on image data obtained by capturing an image of a product. The image inspection device 100 performs a parallel combination of defective product judgment, which judges whether the image data to be inspected is good or defective, and defective product judgment, which judges whether the image data is normal or abnormal. The image inspection device 100 then outputs an inspection result in which the product is good and normal as OK (pass) and in which the product is defective or abnormal as NG (pass).

[0017] Defective product judgment is based on pass / fail rules learned from images of good and defective products collected in advance. When an unknown defective product image is input as an image to be inspected during the learning process, there is a possibility that an image that should be judged as defective will be erroneously judged as good.

[0018] In order to suppress erroneous judgment in the defective product judgment, it is preferable that the defective product judgment is performed in parallel in combination with the defective product judgment. The defective product judgment includes a process of learning the non-defectiveness as a feature from only the good product images, and then calculating the defective product judgment score for the image data of the inspection target. The defective product judgment is performed based on whether the defective product judgment score exceeds a threshold value. The defective product determination score is normalized to a certain range. A specific method for calculating the defective product determination score may be, for example, an algorithm using an autoencoder or an algorithm using a combination of kernel density estimation and a projected histogram.

[0019] In this way, the image inspection device 100 performs defective product judgment and abnormal product judgment in parallel, and can prevent erroneous judgment due to defective product judgment by abnormal product judgment, thereby performing highly accurate inspection. Note that the image inspection device 100 may perform only abnormal product judgment.

[0020] Here, in the case of an algorithm using an autoencoder, the abnormal product judgment score for a normal image may not be stable. In addition, when the image data to be inspected is a color image, the abnormal product judgment score may not be clearly separated between normal images and abnormal images, and the accuracy of the abnormal product judgment may not be achieved. In addition, when using a method that combines kernel density estimation and projected histograms, it is necessary to properly normalize the abnormal product judgment score.

[0021] Therefore, the image inspection device 100 includes a learning dataset acquisition unit 10 that acquires a learning dataset including multiple learning image data, a learning feature derivation unit 20 that derives learning features of the learning image data, a clustering processing unit 30 that performs clustering of the learning image data based on the learning features, an index derivation unit 40 that derives, for each cluster obtained by clustering, the center of gravity of the multiple learning features belonging to the cluster and the distance between the center of gravity and the learning features, and derives the average μ and variance σ of the multiple distances corresponding to the multiple learning features, an imaging feature derivation unit 50 that extracts imaging features based on imaging data captured by a camera C, a minimum distance derivation unit 60 that derives each distance between the imaging features and the multiple centers of gravity and derives a minimum distance X that is the smallest of the distances, an abnormality calculation unit 70 that calculates an abnormality degree α based on the minimum distance X, the average μ, and the variance σ, and a comparison unit 80 that compares the abnormality degree α with a preset threshold. As a result, even if the product to be inspected has multiple variations or the degree of variability in pass / fail products differs depending on the type of product, the abnormality degree α, which is an appropriately normalized evaluation score, is calculated, so that pass / fail judgment of the product to be inspected can be performed based on a stable evaluation score with a clear threshold value between pass / fail products. Therefore, an image inspection device 100 with high pass / fail judgment accuracy can be provided.

[0022] Each of the above-mentioned parts in the image inspection device 100 can be translated as a module or a unit.

[0023] The degree of abnormality α is calculated using the following formula: α={(X-μ) 2 / σ}. This allows the degree of anomaly α to be calculated as an appropriately normalized evaluation score, taking into account the minimum distance X, the mean μ, and the variance σ. This allows the degree of anomaly α, which is an appropriately normalized evaluation score, to be calculated. This allows the degree of anomaly α seen by the user to be always on the same scale, while absorbing the fact that the degree of variation in the appearance of the product to be inspected differs depending on the type of product.

[0024] The image inspection device 100 may include an error processing unit 90 that executes error processing when the degree of abnormality α exceeds a threshold value. The error processing is, for example, displaying on a display that the inspection result of the inspected product is NG (abnormal or defective). This notifies the user that the inspection result of the product is NG.

[0025] The clustering processing unit 30 may perform clustering multiple times using random numbers. Then, the index derivation unit 40 may derive a center of gravity for each cluster each time clustering is performed multiple times, and use a representative cluster that has a minimum residual sum of squares of distances in the cluster as a population for calculating the minimum distance X. This makes it possible to derive a center of gravity that best captures the characteristics of the learning dataset even if the results change each time clustering is tried, thereby providing an image inspection device 100 with higher pass / fail judgment accuracy.

[0026] (Image inspection method) Next, the flow of the image inspection method will be described. The image inspection method can be performed using the image inspection device 100. Each step in the image inspection method can be executed by a corresponding unit in the image inspection device 100. Fig. 3 is a diagram for explaining the flow of the image inspection method.

[0027] (1) As shown in FIG. 3, first, a training dataset including a plurality of training image data is acquired (training dataset acquisition step). The training image data is an image of a non-defective product. That is, the training dataset may be composed of only non-defective product images. The training dataset acquisition step can be executed by the training dataset acquisition unit 10 in the image inspection device 100.

[0028] (2) Next, the learning feature of the learning image data acquired in the learning dataset acquisition step (process) is derived (learning feature derivation step). The learning feature may be derived by an appropriate algorithm based on the learning image data. The algorithm for deriving the learning feature may be, for example, a method (Haar-Like) that uses a set of brightness differences in an image as a feature, a method (HOG) that uses a feature focusing on the distribution of the gradient direction of brightness in a grayscale image, a method (SIFT) that extracts features in two steps of detecting feature points and describing features, a filter method, a wrapper method, or the like. The learning feature derivation step can be executed by the learning feature derivation unit 20 in the image inspection device 100.

[0029] Here, the learning feature derivation step may be executed according to the following flow. Note that an imaging feature derivation step described later may also be executed according to a flow similar to that of the learning feature derivation step shown here. Fig. 4 is a diagram for explaining the flow of the feature derivation step.

[0030] (2-1) As shown in Fig. 4, a resizing process may be performed on image data (learning image data or captured image data). The resizing process is an operation for changing the size of image data, which is performed to unify the input size of a machine learning model.

[0031] (2-2) Next, the image data (learning image data or captured image data) that has been appropriately resized may be padded. The padding process is a process of adding meaningless data before or after short data to adjust its length in order to treat the data as a fixed length.

[0032] (2-3) Next, a window function process may be performed on the image data (learning image data or captured image data) that has been appropriately padded. The window function process is a process of multiplying the image data by a window function in order to extract only information within a finite interval.

[0033] (2-4) Next, FFT processing may be performed on the image data (learning image data or captured image data) that has been appropriately window function processed. The FFT processing is a process of Fourier transforming the image data to convert the spatial domain into the spatial frequency domain based on the image data and obtain the amplitude spectrum distribution.

[0034] (2-5) Next, band-pass filter processing may be performed on the image data (learning image data or captured image data) that has been appropriately FFT-processed. Band-pass filter processing is a process that compresses image data by passing only frequency components in a specific range, thereby removing components that have little effect on the judgment.

[0035] (3) Returning to FIG. 3, next, based on the learning feature derived in the learning feature derivation step, the learning image data is clustered and divided into a plurality of clusters (clustering processing step). Clustering is a machine learning technique that groups data based on the similarity between data. Clustering may be hierarchical clustering using algorithms such as the shortest distance method, the longest distance method, the average distance method, the center of gravity method, and Ward's method, or may be non-hierarchical clustering represented by the k-means method. The algorithm used for clustering may be selected using random numbers each time clustering processing is performed. The algorithm used for clustering may be changed using random numbers each time clustering processing is performed. Specifically, for example, one algorithm may be determined, and the random numbers used therein may be reassigned to perform clustering multiple times. The clustering processing step can be executed by the clustering processing unit 30 in the image inspection device 100.

[0036] (4) Next, for each cluster divided in the clustering process step, the center of gravity of the multiple learning features belonging to the cluster and the distance between the center of gravity and each learning feature are derived. Then, the average μ and variance σ of the multiple distances derived for the multiple learning features are derived (index derivation step). The index derivation step can be executed by the index derivation unit 40 in the image inspection device 100. Through the above steps, a standard index of a non-defective product can be derived as a pre-processing step for the inspection.

[0037] (5) Next, imaging features are extracted based on imaging data captured by the camera (imaging feature derivation step). The imaging data is image data obtained by capturing an image of the product to be inspected by a camera. The imaging features may be derived based on the imaging data using an appropriate algorithm. The imaging features may be extracted using the same method as that used to derive the learning features in the learning feature derivation step. The imaging feature derivation step can be executed by the imaging feature derivation unit 50 in the image inspection device 100.

[0038] (6) Next, the distances between the imaging feature and the multiple centers of gravity are derived, and the minimum distance X that is the smallest among the distances is derived (minimum distance derivation step). The minimum distance derivation step can be executed by minimum distance derivation unit 60 in image inspection device 100. (7) Next, based on the minimum distance X derived in the minimum distance derivation step, and the mean μ and variance σ derived in the index derivation step, specifically, the formula (α={(X-μ) 2 / σ}) (abnormality degree calculation step). The abnormality degree calculation step can be executed by the abnormality degree calculation section 70 in the image inspection device 100.

[0039] (8) Next, the degree of abnormality α calculated in the degree of abnormality calculation step is compared with a preset threshold value (comparison step). The comparison step can be executed by the comparison unit 80 in the image inspection device 100.

[0040] As described above, the image inspection method includes the above-mentioned steps. This makes it possible to stabilize the judgment score and clearly distinguish between good and bad products by a specific threshold value. Therefore, it is possible to improve the accuracy of the final pass / fail judgment of the product.

[0041] (9) Furthermore, if necessary, after the comparison step, error processing may be executed if the degree of abnormality α exceeds a threshold value (error processing step). The error processing is, for example, displaying on the display D a message indicating that the inspection result of the inspected product is NG (abnormal or defective). This notifies the user that the inspection result of the product is NG. The error processing step can be executed by the error processing unit 90 in the image inspection device 100.

[0042] (Image Inspection Program) The image inspection program has a function of causing a computer to execute each process performed at each step in the above-mentioned image inspection method. Specifically, the image inspection program has a learning dataset acquisition function, a learning feature derivation function, a clustering processing function, an index derivation function, an imaging feature derivation function, a minimum distance derivation function, an anomaly degree calculation function, and a comparison function, corresponding to each step in the above-mentioned image inspection method. This makes it possible to stabilize the judgment score and to clearly distinguish between good and bad products at a specific threshold value. This makes it possible to improve the accuracy of the final pass / fail judgment of the product.

[0043] The technical scope of the present invention is not limited to the above-described embodiment, and various modifications can be made without departing from the spirit of the present invention.

[0044] In addition, the components in the above-described embodiment may be replaced with known components as appropriate without departing from the spirit of the present invention. In addition, the above-described modified examples may be combined as appropriate without departing from the spirit of the present invention.

[0045] As described above, the image inspection device 100 according to the embodiment includes a learning dataset acquisition unit 10 that acquires a learning dataset including a plurality of learning image data, a learning feature derivation unit 20 that derives learning features of the learning image data, a clustering processing unit 30 that performs clustering of the learning image data based on the learning features, an index derivation unit 40 that derives, for each cluster obtained by clustering, the center of gravity of the plurality of learning features belonging to the cluster and the distance between the center of gravity and the learning features, and derives the average μ and variance σ of the plurality of distances corresponding to the plurality of learning features, an imaging feature derivation unit 50 that extracts imaging features based on imaging data captured by the camera C, a minimum distance derivation unit 60 that derives each distance between the imaging features and the plurality of centers of gravity and derives the minimum distance X that is the smallest of the distances, an abnormality calculation unit 70 that calculates an abnormality degree α based on the minimum distance X, the average μ, and the variance σ, and a comparison unit 80 that compares the abnormality degree α with a preset threshold value. This allows the calculation of the degree of anomaly α, which is an appropriately normalized evaluation score, so that pass / fail judgment of the product to be inspected can be performed based on a stable evaluation score with a clear threshold value between good and defective products. Therefore, it is possible to provide an image inspection device 100 with high pass / fail judgment accuracy.

[0046] The image inspection method according to the embodiment includes a learning dataset acquisition step of acquiring a learning dataset including a plurality of learning image data, a learning feature derivation step of deriving learning features of the learning image data, a clustering process step of clustering the learning image data based on the learning features to divide the learning image data into a plurality of clusters, an index derivation step of deriving the center of gravity of the plurality of learning features belonging to each cluster and the distance between the center of gravity and the learning features for each cluster, and deriving the average μ and variance σ of the plurality of distances corresponding to the plurality of learning features, an imaging feature derivation step of extracting imaging features based on imaging data captured by a camera, a minimum distance derivation step of deriving each distance between the imaging features and the plurality of centers of gravity and deriving the minimum distance X that is the smallest of the distances, an anomaly degree calculation step of calculating an anomaly degree α based on the minimum distance X, the average μ and the variance σ, and a comparison step of comparing the anomaly degree α with a preset threshold value. As a result, the anomaly degree α, which is an appropriately normalized evaluation score, is calculated, so that the pass / fail judgment of the product to be inspected can be performed based on a stable evaluation score with a clear threshold value between a good product and a defective product. Therefore, an image inspection method with high pass / fail judgment accuracy can be provided.

[0047] The image inspection program according to the embodiment causes a computer to execute a learning dataset acquisition function for acquiring a learning dataset including a plurality of learning image data, a learning feature derivation function for deriving learning features of the learning image data, a clustering processing function for clustering the learning image data based on the learning features to divide the learning image data into a plurality of clusters, an index derivation function for deriving the center of gravity of the plurality of learning features belonging to each cluster and the distance between the center of gravity and the learning features for each cluster, and deriving the average μ and variance σ of the plurality of distances corresponding to the plurality of learning features, an imaging feature derivation function for extracting imaging features based on imaging data captured by a camera, a minimum distance derivation function for deriving each distance between the imaging features and the plurality of centers of gravity and deriving the minimum distance X that is the smallest of the respective distances, an anomaly degree calculation function for calculating an anomaly degree α based on the minimum distance X, the average μ and the variance σ, and a comparison function for comparing the anomaly degree α with a preset threshold value. As a result, the anomaly degree α, which is an appropriately normalized evaluation score, is calculated, so that the pass / fail judgment of the product to be inspected can be performed based on a stable evaluation score with a clear threshold value between a good product and a defective product. Therefore, an image inspection program with high pass / fail judgment accuracy can be provided. [Explanation of symbols]

[0048] 100 Image inspection device 10. Learning Data Set Acquisition Unit 20 Learning feature extraction unit 30 Clustering Processing Unit 40 Indicator derivation part 50 Imaging feature extraction unit 60 Minimum distance derivation part 70 Abnormality degree calculation section 80 Comparison section 90 Error processing section α Abnormality μ average σ variance B Bus C Camera D Display K Keyboard M Memory S Control Means X minimum distance

Claims

1. A learning dataset acquisition unit that acquires a learning dataset including a plurality of learning image data; A learning feature derivation unit that derives learning features of the learning image data; a clustering processing unit that performs clustering on the learning image data based on the learning feature amount and divides the learning image data into a plurality of clusters; an index derivation unit that derives, for each cluster, a center of gravity of the plurality of learning features belonging to the cluster and a distance between the center of gravity and the learning features, and derives an average μ and a variance σ of the plurality of distances corresponding to the plurality of learning features; an imaging feature derivation unit that extracts imaging feature amounts based on imaging data captured by a camera; a minimum distance deriving unit that derives distances between the imaging feature amount and a plurality of the centers of gravity and derives a minimum distance X that is the smallest of the distances; an anomaly level calculation unit that calculates an anomaly level α based on a minimum distance X, the mean μ, and the variance σ; A comparison unit that compares the abnormality degree α with a preset threshold value; An image inspection device comprising:

2. An error processing unit that executes an error process when the abnormality degree α exceeds the threshold value.

2. The image inspection device according to claim 1.

3. The clustering processing unit performs the clustering process a plurality of times using random numbers; The index derivation unit derives the center of gravity for each cluster each time the clustering is performed, and determines a representative cluster for which a residual sum of squares of the distances in the clusters is minimum as a population for calculating the minimum distance X.

3. An image inspection device according to claim 1 or 2.

4. The imaging feature amount derivation unit or the learning feature amount derivation unit performs resizing processing, window function processing, FFT processing, or band pass filter processing.

3. An image inspection device according to claim 1 or 2.

5. A learning dataset acquisition step of acquiring a learning dataset including a plurality of learning image data; A learning feature derivation step of deriving learning features of the learning image data; A clustering process step of clustering the learning image data based on the learning feature amount and dividing the learning image data into a plurality of clusters; an index derivation step of deriving, for each cluster, a center of gravity of the plurality of learning features belonging to the cluster and a distance between the center of gravity and the learning feature, and deriving an average μ and a variance σ of the plurality of distances corresponding to the plurality of learning features; An imaging feature amount derivation step of extracting imaging feature amounts based on imaging data captured by a camera; a minimum distance deriving step of deriving each distance between the imaging feature amount and a plurality of the centers of gravity, and deriving a minimum distance X that is the smallest among the distances; an anomaly calculation step of calculating an anomaly degree α based on a minimum distance X, the mean μ, and the variance σ; a comparison step of comparing the abnormality degree α with a preset threshold value; An imaging inspection method comprising:

6. On the computer, A learning dataset acquisition function for acquiring a learning dataset including a plurality of learning image data; A learning feature derivation function for deriving learning features of the learning image data; A clustering processing function for clustering the learning image data based on the learning feature amount and dividing the learning image data into a plurality of clusters; an index derivation function for deriving, for each cluster, a center of gravity of the plurality of learning features belonging to the cluster and a distance between the center of gravity and the learning feature, and deriving an average μ and a variance σ of the plurality of distances corresponding to the plurality of learning features; An imaging feature derivation function that extracts imaging feature amounts based on imaging data captured by a camera; a minimum distance deriving function of deriving each distance between the imaging feature amount and a plurality of the centers of gravity and deriving a minimum distance X that is the smallest among the distances; an anomaly degree calculation function that calculates an anomaly degree α based on a minimum distance X, the mean μ, and the variance σ; A comparison function for comparing the abnormality degree α with a preset threshold value; An image inspection program that executes the above.

Images