Image processing device and image processing system

Abstract

To provide an image processing device and an image processing system, capable of suppressing the reduction of processing accuracy.SOLUTION: An X-ray inspection device 3 includes an acquisition part 21 for acquiring a first parameter related to image pickup when picking up an image used for machine learning, a setting part 22 for setting a second parameter related to the image pickup when picking up an object image which is subjected to image processing, and a comparison part 23 for comparing the first parameter acquired by the acquisition part 21 with the second parameter set by the setting part 22.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to an image processing device and an image processing system. [Background technology]

[0002] Known conventional image processing devices include, for example, the device described in Patent Document 1. The device described in Patent Document 1 is an inspection device that inspects an item by image processing, and includes an X-ray irradiation unit that irradiates the item with X-rays and an X-ray detection unit that detects the X-rays, and inspects the item by performing image processing on an X-ray image created based on the X-ray detection results. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] International Publication No. 2006 / 001107 Summary of the Invention [Problem to be solved by the invention]

[0004] In recent years, inspection devices that inspect items using image processing have been using trained models generated by machine learning to perform foreign substance inspections, etc. In image processing using a trained model, it is necessary for predetermined parameters (such as settings of imaging equipment and parameters related to image processing methods) to match between first parameters related to image capture when capturing an image used in machine learning of the trained model and second parameters related to image capture when capturing a target image to be processed. Typically, the second parameters are manually set by an operator or the like. If the second parameters, once set, are changed for some reason after image processing has begun, the image processing will be performed with the two parameters in a different state. In this case, the accuracy of image processing using the trained model will decrease.

[0005] An object of one aspect of the present invention is to provide an image processing device and an image processing system that can suppress a decrease in processing accuracy. [Means for solving the problem]

[0006] An image processing device according to one aspect of the present invention is an image processing device that performs image processing using a trained model generated by machine learning using an image including an object, and includes an acquisition unit that acquires first parameters related to image capture when capturing an image used in machine learning, a setting unit that sets second parameters related to image capture when capturing a target image that is the subject of image processing, and a comparison unit that compares the first parameters acquired by the acquisition unit with the second parameters set by the setting unit.

[0007] In an image processing device according to one aspect of the present invention, a comparison unit compares a first parameter acquired by an acquisition unit with a second parameter set by a setting unit. This allows the image processing device to recognize that the two parameters are different when the first parameter and the second parameter are different. Therefore, when the two parameters are different, the image processing device can notify the user that the two parameters are different or adjust the second parameter, thereby avoiding image processing when the two parameters are different. Therefore, the image processing device can suppress a decrease in processing accuracy.

[0008] In one embodiment, when the comparison between the first parameter and the second parameter by the comparison unit reveals that the first parameter and the second parameter have different predetermined parameters, the setting unit may change the setting of the second parameter based on the first parameter. In this configuration, even if the first parameter and the second parameter have different predetermined parameters (parameters that must match), the setting unit automatically changes the second parameter. Therefore, the image processing device can suppress a decrease in processing accuracy.

[0009] In one embodiment, the image processing device may include a notification unit that notifies the user that the second parameter is different when a predetermined parameter is different between the first parameter and the second parameter as a result of the comparison between the first parameter and the second parameter by the comparison unit. In this configuration, the notification unit notifies the user that the second parameter is different, so that the user can be notified that the second parameter is different. This allows the user to change the second parameter, thereby preventing a decrease in processing accuracy.

[0010] In one embodiment, the image processing device may include a stop unit that stops the execution of image processing when a predetermined parameter differs between the first parameter and the second parameter as a result of the comparison between the first parameter and the second parameter by the comparison unit. In this configuration, if the predetermined parameter differs between the first parameter and the second parameter, the image processing is not performed. Therefore, the image processing device can suppress a decrease in processing accuracy.

[0011] In one embodiment, the target image may include an object, and an inspection unit may be provided that, when a predetermined parameter is found to match between the first parameter and the second parameter as a result of the comparison between the comparison unit and the second parameter, inspects the object based on the target image. This configuration can suppress a decrease in processing accuracy with respect to the inspection of the object. In other words, when inspecting an object using an image processing device, a decrease in inspection accuracy can be suppressed.

[0012] An image processing system according to one aspect of the present invention includes the image processing device described above, an image acquisition device that captures an image and outputs the image in association with a first parameter related to image capture at the time the image was captured, and a machine learning device that generates a trained model by machine learning using the image output from the image acquisition device, and outputs the trained model in association with the first parameter to the image processing device.

[0013] An image processing system according to one aspect of the present invention includes the image processing device described above, and therefore, the image processing system can suppress a decrease in processing accuracy. [Effects of the Invention]

[0014] According to one aspect of the present invention, it is possible to suppress a decrease in processing accuracy. [Brief explanation of the drawings]

[0015] [Figure 1] FIG. 1 is a diagram illustrating an inspection system according to an embodiment. [Figure 2] FIG. 2 is a diagram showing an X-ray inspection device. [Figure 3] FIG. 3 is a diagram showing the internal configuration of the shielding box shown in FIG. [Figure 4] FIG. 4 is a diagram illustrating the configuration of the control unit. [Figure 5] FIG. 5 is a diagram illustrating a neural network. DETAILED DESCRIPTION OF THE INVENTION

[0016] Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the description of the drawings, the same or corresponding elements are designated by the same reference numerals, and redundant description will be omitted.

[0017] As shown in FIG. 1, the inspection system (image processing system) 1 includes an image acquisition device 2, an X-ray inspection device 3, and a machine learning device 5.

[0018] The image acquisition device 2 acquires an image. The image acquisition device 2 is an X-ray inspection device. The image acquisition device 2 has a configuration similar to that of the X-ray inspection device 3. The image acquisition device 2 captures an image and stores the image in association with a first parameter related to image capture at the time the image was captured. The image acquisition device 2 outputs image data including the image and the first parameter to the machine learning device 5.

[0019] The first parameters are parameters relating to the settings of the imaging equipment of the image acquisition device 2 and the image processing method. The first parameters include, for example, irradiation unit parameters relating to the settings of the X-ray irradiation unit when an image is captured in the image acquisition device 2, detection unit parameters relating to the X-ray detection unit, and processing parameters relating to image processing. The irradiation unit parameters are, for example, the setting values ​​of the voltage and current of the X-ray tube. The irradiation unit parameters are set as appropriate in the image acquisition device 2. The detection unit parameters are, for example, the pixel size and type (model) of the line sensor. The processing parameters are, for example, the processing content performed on the image (noise removal, contrast adjustment, etc.). The processing parameters are set as appropriate in the image acquisition device 2.

[0020] The machine learning device 5 is a device that generates a trained model through machine learning. The machine learning device 5 is composed of a CPU (Central Processing Unit), ROM (Read Only Memory), RAM (Random Access Memory), etc. The image acquisition device 2 and the X-ray inspection device 3 are communicably connected to the machine learning device 5 via a wired or wireless network such as the Internet or a telephone network, and can send and receive information.

[0021] The machine learning device 5 includes a communication unit 30, a learning model generation unit 31, and a storage unit 32.

[0022] The communication unit 30 communicates with the image acquisition device 2 and the X-ray inspection device 3. The communication unit 30 receives image data transmitted from the image acquisition device 2 and outputs it to the learning model generation unit 31 and the storage unit 32. The communication unit 30 transmits data of the trained model generated in the learning model generation unit 31 (data in "xxx.h5") and data indicating the first parameter to the X-ray inspection device 3. The data of the first parameter is, for example, text data (data in "xxx.txt"). Note that the first parameter may be included in the file name of the trained model data (xxx(first parameter).h5).

[0023] The learning model generation unit 31 acquires training data to be used for machine learning, and performs machine learning using the acquired training data to generate a trained model. The training data includes images and other data. The images are images acquired by the image acquisition device 2. The other data includes item information indicating the item in the image. The other data also includes first parameters related to image capture when the image is captured.

[0024] The learning model generation unit 31 performs machine learning using each pixel value of the image as an input value to the neural network and processing information corresponding to the image as an output value of the neural network to generate (configure) a neural network NW (see FIG. 5). When pixel values ​​are used as input values, they are used as input values ​​for neurons associated with each pixel (the position of the pixel on the image). The above machine learning itself can be performed in the same way as conventional machine learning algorithms. The learning model generation unit 31 stores the generated trained model in the storage unit 32. The learning model generation unit 31 associates the data of the trained model with parameter information including the first parameter and stores it in the storage unit 32.

[0025] As shown in FIG. 2, the X-ray inspection device 3 includes an apparatus main body 10, support legs 11, a shielding box 12, a conveying unit 13, an X-ray irradiation unit 14, an X-ray detection unit 15, a display operation unit 16, and a control device 17. The X-ray inspection device 3 generates an X-ray transmission image of the object G while conveying the object G, and performs inspection of the object G (e.g., foreign body inspection, storage quantity inspection, missing item inspection, crack / chip inspection, etc.) based on the X-ray transmission image. The object G before inspection is carried into the X-ray inspection device 3 by a carry-in conveyor 51. The object G after inspection is carried out from the X-ray inspection device 3 by an unloading conveyor 52. The object G determined to be a defective item by the X-ray inspection device 3 is sorted out of the production line by a sorting device (not shown) arranged downstream of the unloading conveyor 52. The object G determined to be a non-defective item by the X-ray inspection device 3 passes directly through the sorting device.

[0026] The device main body 10 houses the control device 17 and the like. The support legs 11 support the device main body 10. The shielding box 12 is provided on the device main body 10. The shielding box 12 prevents leakage of X-rays (electromagnetic waves) to the outside. An inspection area R is provided inside the shielding box 12, where inspection of items G is carried out using X-rays. The shielding box 12 is formed with an inlet 12a and an outlet 12b. Items G before inspection are carried into the inspection area R from the carry-in conveyor 51 via the inlet 12a. Items G after inspection are carried out from the inspection area R to the outlet 12b to the outlet conveyor 52. Each of the inlet 12a and the outlet 12b is provided with an X-ray shielding curtain (not shown) that prevents leakage of X-rays.

[0027] The transport unit 13 is disposed so as to penetrate the center of the shielding box 12. The transport unit 13 transports the article G along the transport direction A from the entrance 12a through the inspection area R to the exit 12b. The transport unit 13 is, for example, a belt conveyor stretched between the entrance 12a and the exit 12b. Note that the transport unit 13, which is a belt conveyor, may protrude outward beyond the entrance 12a and the exit 12b.

[0028] As shown in Figures 2 and 3, the X-ray irradiator 14 is disposed within the shield box 12. The X-ray irradiator 14 irradiates X-rays onto the object G transported by the transport unit 13. The X-ray irradiator 14 includes, for example, an X-ray tube that emits X-rays and a diaphragm that spreads the X-rays emitted from the X-ray tube in a fan shape in a plane perpendicular to the transport direction A. The X-rays irradiated from the X-ray irradiator 14 include X-rays in various energy bands ranging from low energy (long wavelength) to high energy (short wavelength). Note that the terms "low" and "high" in the low energy band and high energy band described above indicate relatively "low" and "high" among the multiple energy bands irradiated from the X-ray irradiator 14, and do not indicate a specific range.

[0029] The X-ray detection unit 15 is disposed within the shielding box 12. The X-ray detection unit 15 detects X-rays in each of a plurality of energy bands that have passed through the article G. In this embodiment, the X-ray detection unit 15 is configured to detect X-rays in a low-energy band and X-rays in a high-energy band. That is, the X-ray detection unit 15 has a first line sensor 18 and a second line sensor 19. The first line sensor 18 and the second line sensor 19 are each configured with X-ray detection elements that are linearly arranged along a horizontal direction perpendicular to the conveying direction A. The first line sensor 18 detects X-rays in the low-energy band that have passed through the article G and the conveying belt of the conveying unit 13. The second line sensor 19 detects X-rays in the high-energy band that have passed through the article G, the conveying belt of the conveying unit 13, and the first line sensor 18.

[0030] As shown in Fig. 2, the display operation unit 16 is provided on the device main body 10. The display operation unit 16 displays various information and accepts input of various conditions. The display operation unit 16 is, for example, a liquid crystal display, and displays an operation screen as a touch panel. In this case, the operator can input various conditions via the display operation unit 16.

[0031] The control device 17 is disposed within the device body 10. The control device 17 controls the operation of each part of the X-ray inspection device 3 (in this embodiment, the conveying unit 13, the X-ray irradiation unit 14, the X-ray detection unit 15, and the display operation unit 16, as well as a sorting device (not shown) disposed downstream of the X-ray inspection device 3). The sorting device is a device that removes from the conveying path an item G determined to be a defective item by image inspection by the X-ray inspection device 3. The control device 17 includes a processor such as a CPU (Central Processing Unit), memories such as a ROM (Read Only Memory) and a RAM (Random Access Memory), and storage such as an SSD (Solid State Drive). A program for controlling the X-ray inspection device 3 is recorded in the ROM. The control device 17 receives an input of a detection result of low-energy band X-rays from a first line sensor 18 of the X-ray detection unit 15, and an input of a detection result of high-energy band X-rays from a second line sensor 19 of the X-ray detection unit 15.

[0032] As shown in FIG. 4, the control device 17 includes a communication unit 20, an acquisition unit 21, a setting unit 22, a comparison unit 23, and an inspection unit 24.

[0033] The communication unit 20 communicates with the machine learning device 5. The communication unit 20 receives data and parameter information of the trained model transmitted from the machine learning device 5. The communication unit 20 outputs the trained model to a storage unit (not shown) and outputs the parameter information to the acquisition unit 21.

[0034] The acquisition unit 21 acquires first parameters related to image capture when an image used in machine learning is captured. The acquisition unit 21 acquires the first parameters from parameter information output from the communication unit 20. Based on the parameter information, the acquisition unit 21 acquires, as the first parameters, irradiation unit parameters related to the settings of the X-ray irradiation unit, information related to the X-ray detection unit, and processing information related to image processing. The acquisition unit 21 outputs the acquired first parameters to the comparison unit 23.

[0035] The setting unit 22 sets second parameters related to image capture when capturing a target image to be inspected (image processing). As the second parameters, the setting unit 22 sets irradiation unit parameters related to setting the X-ray irradiation unit 14, detection unit parameters related to the X-ray detection unit 15, and processing parameters related to image processing. The irradiation unit parameters and processing parameters are set by, for example, the display operation unit 16. Note that the irradiation unit parameters and processing parameters may be acquired from another device via the communication unit 20.

[0036] The setting unit 22 sets the second parameter based on the comparison result described below. When the comparison between the first parameter and the second parameter by the comparison unit 23 shows that the first parameter and the second parameter have different predetermined parameters, the setting unit 22 automatically changes the setting of the second parameter based on the first parameter. The setting unit 22 changes (rewrites) the second parameter of the X-ray inspection apparatus 3 to the first parameter. When the setting unit 22 sets the second parameter, it causes the display operation unit 16 to display that effect.

[0037] The comparison unit 23 compares the first parameter acquired by the acquisition unit 21 with the second parameter set by the setting unit 22. The comparison unit 23 compares the first parameter with the second parameter at a predetermined timing. The predetermined timing is, for example, when the X-ray inspection apparatus 3 is powered on. The comparison unit 23 compares the first parameter acquired by the acquisition unit 21 with the second parameter set by the setting unit 22, and determines whether the predetermined parameters match. The predetermined parameters are, for example, among the irradiation unit parameters, detection unit parameters, and processing parameters, the voltage of the X-ray tube of the X-ray irradiation unit 14, the pixel size and type (model) of the line sensor of the X-ray detection unit 15, and parameters related to contrast adjustment in image processing. The predetermined parameters are parameters that must match and are set appropriately. The comparison unit 23 outputs the comparison result between the first parameter and the second parameter to the setting unit 22.

[0038] If the comparison unit 23 compares the first parameter with the second parameter and finds that the predetermined parameters in the first parameter and the second parameter match, the inspection unit 24 inspects the item G based on the inspection image (target image). The inspection unit 24 generates an inspection image based on the detection results of the first line sensor 18 and the second line sensor 19. The control device 17 generates a soft image based on the detection results of the low-energy band X-rays by the first line sensor 18. The control device 17 generates a hard image based on the detection results of the high-energy band X-rays by the second line sensor 19. The control device 17 generates an inspection image based on the soft image and the hard image.

[0039] The inspection unit 24 inspects the presence or absence of a foreign object in the item G using the trained model. The trained model predicts and outputs the presence or absence of a foreign object in the item G shown in an image based on the image data. The trained model includes a neural network NW. The trained model may include a convolutional neural network. Furthermore, the trained model may include a neural network with multiple layers (e.g., eight or more layers). In other words, the trained model may be generated by deep learning.

[0040] As shown in FIG. 5, the neural network NW is composed of, for example, a first layer which is an input layer, a second layer, a third layer, and a fourth layer which are intermediate layers (hidden layers), and a fifth layer which is an output layer. The first layer receives an input value x=(x0, x1, x2, ... x) with p parameters as elements. p ) is output to the second layer as is. The second, third, and fourth layers each use an activation function to convert the total input into an output, and pass that output on to the next layer. The fifth layer also uses an activation function to convert the total input into an output, which is the output value y = (y0) of a neural network with one parameter as an element.

[0041] In this embodiment, the neural network NW receives the pixel values ​​of each pixel in the inspection image and outputs information indicating the presence or absence of a foreign substance. The input layer of the neural network NW is provided with neurons equal to the number of pixels in the inspection image. The output layer of the neural network NW is provided with neurons for outputting information indicating the presence or absence of a foreign substance. The information indicating the presence or absence of a foreign substance can be determined based on the output values ​​of the neurons in the output layer. The output value of the neuron is, for example, a value between 0 and 1. In this case, a larger neuron value (closer to 1) indicates that the item G contains a foreign substance (is abnormal), and a smaller neuron value (closer to 0) indicates that the item G does not contain a foreign substance (is normal).

[0042] The inspection unit 24 inputs an inspection image into the trained model. The inspection unit 24 may normalize the inspection image to be input. The normalization of the inspection image is performed, for example, by reducing, enlarging, or trimming the image. The inspection unit 24 may also perform various processes on the inspection image to be input, such as adjusting the contrast, changing the color, or changing the format. In response to inputting the inspection image into the neural network NW of the trained model, the inspection unit 24 obtains an inspection result including an output value output from the neural network NW.

[0043] The inspection unit 24 determines whether the item G is good or bad based on the inspection results. If the inspection unit 24 determines that there is a foreign object in the item G, it determines that the item G is defective. If the inspection unit 24 determines that the item G is defective, it causes the display operation unit 16 to display that fact.

[0044] As described above, in the X-ray inspection apparatus 3 according to this embodiment, the comparison unit 23 compares the first parameter acquired by the acquisition unit 21 with the second parameter set by the setting unit 22. As a result, when the first parameter and the second parameter are different, the X-ray inspection apparatus 3 can grasp that the two parameters are different. Therefore, when the two parameters are different, the X-ray inspection apparatus 3 can notify the difference or adjust the second parameter, thereby preventing the inspection from being performed in a state where the two parameters are different. Therefore, the X-ray inspection apparatus 3 can suppress a decrease in processing accuracy.

[0045] In the X-ray inspection apparatus 3 according to this embodiment, when a predetermined parameter differs between the first parameter and the second parameter as a result of the comparison between the first parameter and the second parameter by the comparison unit 23, the setting unit 22 changes the setting of the second parameter based on the first parameter. In this configuration, even if a predetermined parameter (a parameter that must match) differs between the first parameter and the second parameter, the setting unit 22 automatically changes the second parameter. Therefore, the X-ray inspection apparatus 3 can suppress a decrease in processing accuracy.

[0046] The X-ray inspection apparatus 3 according to this embodiment includes an inspection unit 24 that inspects the article G based on the inspection image when predetermined parameters match between the first parameter and the second parameter as a result of the comparison between the first parameter and the second parameter by the comparison unit 23. This configuration can suppress a decrease in processing accuracy with respect to the inspection of the article G. In other words, when inspecting the article G in the X-ray inspection apparatus 3, a decrease in inspection accuracy can be suppressed.

[0047] Although the embodiments of the present invention have been described above, the present invention is not necessarily limited to the above-described embodiments, and various modifications are possible without departing from the spirit of the present invention.

[0048] In the above embodiment, an example has been described in which the image processing system is the inspection system 1 and the image processing device is the X-ray inspection device 3. However, the image processing device is not limited to the X-ray inspection device. For example, the image processing device may be a device that inspects an item based on an image captured by a camera, or a device that inspects an item by irradiating it with near-infrared light.

[0049] In the above embodiment, an example has been described in which the machine learning device 5 generates a trained model through machine learning, and the X-ray inspection device 3 performs processing using the trained model. However, the processing using the trained model may be performed by the machine learning device 5. That is, the machine learning device 5 may generate a trained model through machine learning and perform processing using the trained model. In this configuration, the processing results using the trained model are transmitted to the X-ray inspection device 3. Furthermore, the inspection system 1 may be configured with a single X-ray inspection device. That is, the image acquisition device, the machine learning device, and the image processing device may be configured with a single device.

[0050] In the above embodiment, an example has been described in which the communication unit 20 of the X-ray inspection apparatus 3 receives data and parameter information of a trained model transmitted from the machine learning apparatus 5. Trained models are portable between computer systems. Therefore, the trained model generated by the machine learning apparatus 5 may be ported to the X-ray inspection apparatus 3 using a recording medium or the like.

[0051] In the above embodiment, an example has been described in which X-ray detection unit 15 in X-ray inspection apparatus 3 has first line sensor 18 and second line sensor 19. However, X-ray detection unit 15 may have only one line sensor.

[0052] In the above embodiment, an example has been described in which the inspection unit 24 inspects the presence or absence of foreign matter in the items G. However, the inspection unit 24 may also inspect the number of items stored, inspect for missing items, inspect for cracks and chips, and the like.

[0053] In the above embodiment, an example has been described in which, when the comparison unit 23 compares the first parameter with the second parameter and finds that the predetermined parameter differs between the first parameter and the second parameter, the setting unit 22 automatically changes the setting of the second parameter based on the first parameter. However, the X-ray inspection apparatus 3 (image processing apparatus) may also include a notification unit that notifies the user that the second parameter differs when the comparison unit 23 compares the first parameter with the second parameter and finds that the predetermined parameter differs between the first parameter and the second parameter. The notification unit may be the display operation unit 16. That is, when the predetermined parameter differs between the first parameter and the second parameter, the display operation unit 16 may display that the second parameter differs. In this configuration, the notification that the second parameter differs can be made to notify the user, etc. that the second parameter differs. This allows the user, etc., to change the second parameter, thereby preventing a decrease in processing accuracy.

[0054] In the above embodiment, the comparison unit 23 compares the first parameter with the second parameter when the X-ray inspection apparatus 3 is powered on. However, the timing at which the comparison unit 23 compares the parameters is not limited to this. For example, the comparison unit 23 may compare the parameters while inspecting the item G. In this case, the X-ray inspection apparatus 3 (image processing apparatus) may include a stop unit that stops the inspection (image processing) when a predetermined parameter differs between the first parameter and the second parameter as a result of the comparison between the first parameter and the second parameter by the comparison unit 23. In this configuration, image processing is not performed when a predetermined parameter differs between the first parameter and the second parameter. Therefore, the X-ray inspection apparatus 3 can suppress a decrease in processing accuracy. [Explanation of symbols]

[0055] 1...inspection system (image processing system), 2...image acquisition device, 3...X-ray inspection device (image processing device), 5...machine learning device, 21...acquisition unit, 22...setting unit, 23...comparison unit, 24...inspection unit, G...item.

Claims

1. An image processing device that performs image processing to inspect whether an item in an image is defective using a trained model generated by machine learning using an image, an acquisition unit that acquires, for the trained model, a first parameter related to image capture when the image used in the machine learning is captured; a setting unit that sets second parameters related to image capture when capturing a target image that is a target of image processing; a comparison unit that compares the first parameter acquired by the acquisition unit with the second parameter set by the setting unit, The setting unit changes the setting of the second parameter based on the first parameter when a comparison between the first parameter and the second parameter by the comparison unit reveals that a predetermined parameter differs between the first parameter and the second parameter.

2. 2. The image processing device according to claim 1, further comprising: an alarm unit that, when a comparison between the first parameter and the second parameter by the comparison unit reveals that a predetermined parameter is different between the first parameter and the second parameter, notifies the user that the second parameter is different.

3. 3. The image processing device according to claim 1, further comprising a stop unit that stops the execution of the image processing when a comparison between the first parameter and the second parameter by the comparison unit reveals that a predetermined parameter differs between the first parameter and the second parameter.

4. the target image includes an item; The image processing device according to any one of claims 1 to 3, further comprising an inspection unit that inspects the item based on the target image when a comparison between the first parameter and the second parameter by the comparison unit shows that a predetermined parameter in the first parameter matches that in the second parameter.

5. The image processing device according to any one of claims 1 to 4, an image acquisition device that captures an image and outputs the image in association with a first parameter related to image capture when the image is captured; An image processing system comprising: a machine learning device that generates a trained model through machine learning using the image output from the image acquisition device, associates the trained model with the first parameter, and outputs the trained model to the image processing device.

Images