JP2023035618A - Anomaly detection device and anomaly detection method - Google Patents

Abstract

To provide an anomaly detection device and anomaly detection method, which enable accurate detection of shape anomalies of target objects.SOLUTION: An anomaly detection device is provided, comprising an acquisition unit configured to acquire a captured image, captured by an image capturing device, of a predetermined object of a part being conveyed on a conveying path, a division unit configured to divide an image portion of the target object in the captured image acquired by the acquisition unit into multiple images, and a determination unit configured to perform anomaly determination processing on an input of the multiple images divided by the division unit using a predetermined learning model and obtain a determination result indicative of the presence or absence of anomalies as an output.SELECTED DRAWING: Figure 3

Description

Applied for application of Article 30, Paragraph 2 of the Patent Act Date of publication August 20, 2021 Address of publication http://gijutsu12/Giken/index. html

The present invention relates to an anomaly detection device and an anomaly detection method.

In a conventional automatic transport production line, imaging devices are installed at key points to perform visual confirmation of parts or image judgment of captured images based on a rule base. As a technique for image determination for images captured by an imaging device in such a production line, an acquisition unit that acquires an image of the surface of an object, and an image of the surface of the object that is obtained from the image of the surface of the object acquired by the acquisition unit. an inspection system comprising: (see, for example, Patent Document 1).

WO2019/112040

However, in conventional manufacturing line technology, image processing based on simple color determination processing, etc. cannot detect shape abnormalities for parts such as automobile parts that have many characteristic curved portions such as unevenness. There is a problem that it is difficult to

SUMMARY OF THE INVENTION It is an object of the present invention to provide an anomaly detection device and an anomaly detection method capable of accurately detecting an anomaly in the shape of an object.

In order to solve the above-described problems and achieve the object, an abnormality detection device according to the present invention acquires a captured image of a predetermined object in a component transported on a transport path, which is imaged by an imaging device. a dividing unit that divides the image portion of the object in the captured image acquired by the acquiring unit into a plurality of images; a determination unit that obtains as an output a determination result indicating the presence or absence of an abnormality by an abnormality determination process using a model.

ADVANTAGE OF THE INVENTION According to this invention, the abnormality of a shape can be detected accurately about a target object.

FIG. 1 is a diagram showing an example of the overall configuration of an anomaly detection system according to an embodiment. FIG. 2 is a diagram illustrating an example of the hardware configuration of the abnormality detection device according to the embodiment; FIG. 3 is a diagram illustrating an example of a functional block configuration of the anomaly detection device according to the embodiment; FIG. 4 is a diagram showing an example of an extracted image including a picked-up image of the roof member and a flange portion. FIG. 5 is a diagram showing an example of a normal image of the flange portion. FIG. 6 is a diagram showing an example of an abnormal image of the flange portion. FIG. 7 is a diagram for explaining a state in which the extracted image of the flange portion is divided. FIG. 8 is a diagram showing an example of a superimposed display of the determination result regarding the shape of the flange portion of the captured image. FIG. 9 is a flowchart showing an example of the flow of anomaly detection processing of the anomaly detection system according to the embodiment.

1 to 9, embodiments of an abnormality detection device and an abnormality detection method according to the present invention will be described in detail below. In addition, the present invention is not limited by the following embodiments, and the constituent elements in the following embodiments can be easily conceived by those skilled in the art, substantially the same, and so-called equivalent ranges. is included. Furthermore, various omissions, replacements, changes and combinations of components can be made without departing from the gist of the following embodiments.

(Overall configuration of anomaly detection system)
FIG. 1 is a diagram showing an example of the overall configuration of an anomaly detection system according to an embodiment. An overall configuration of an anomaly detection system 1 according to the present embodiment will be described with reference to FIG.

An anomaly detection system 1 shown in FIG. 1 is a system for detecting an anomaly in the shape of a predetermined object in a part indicated by an image captured by an imaging device on a transport line 3 (transport path). As shown in FIG. 1, the anomaly detection system 1 includes a PLC (Programmable Logic Controller) 11, an imaging device 12, an anomaly detection device 10, a display device 13, and a server device 20.

The PLC 11 is an industrial control device that controls the installation operation, assembly operation, transport operation, and the like of the parts 30 such as the vehicle body on the transport line 3 . The PLC 11 transmits a transport signal indicating that the component 30 has been transported to a predetermined position on the transport line 3 at a timing suitable for imaging by the imaging device 12 to the abnormality detection device 10. . A carrier signal from the PLC 11 may be sent to the abnormality detection device 10 directly or via a relay component or the like. For example, the PLC 11 controls the timing when the part 30 is placed at a predetermined position suitable for imaging by the imaging device 12 by controlling an articulated robot or the like, or the part 30 is conveyed to the predetermined position by a limiter switch, a proximity sensor, a photoelectric sensor, or the like. A carrier signal is transmitted to the abnormality detection device 10 at the timing when it is detected. The abnormality detection device 10 may receive a transport signal indicating that the component 30 has been transported to a predetermined position from a limiter switch, a proximity sensor, a photoelectric sensor, or the like installed on the transport line 3 .

The image capturing device 12 is an image capturing device that captures an image of a region including the flange portion of the roof member, which is a predetermined target object in the component 30 . The imaging device 12 transmits the captured image to the abnormality detection device 10 . In this embodiment, it is assumed that the component 30 is the vehicle body of an automobile, and the predetermined object is the flange portion of the roof member of the vehicle body. Note that the part 30 and the predetermined object are not limited to these.

The image captured by the imaging device 12 may be still image data or moving image data.

The abnormality detection device 10 is a device for detecting an abnormality in the shape of the flange portion of the roof member of the component 30 included in the captured image captured by the imaging device 12 . The anomaly detection device 10 is realized by, for example, a single board computer. The abnormality detection device 10 is not limited to being a single board computer, and may be an information processing device such as a normal PC (Personal Computer) or workstation. The abnormality detection device 10 includes an image processing program 101 , a preprocessing program 102 and an abnormality determination program 103 .

The image processing program 101 is a program for performing image processing such as extracting an image of the flange portion of the roof member from the captured image captured by the imaging device 12 . Note that the extraction method in the image processing program 101 includes extraction processing of a predetermined region of the captured image, extraction processing based on a predetermined algorithm such as template matching, or YOLO ( It may be a method based on any existing extraction process using You Only Look Once).

The preprocessing program 102 is a program for executing predetermined preprocessing for abnormality determination by the abnormality determination program 103 in the latter stage on the extracted image extracted by the image processing program 101 . The preprocessing program 102 divides the extracted image into a plurality of images as predetermined preprocessing, specifically, in order to reduce the load of abnormality determination processing by the abnormality determination program 103 .

The abnormality determination program 103 is a program for executing abnormality determination processing using a predetermined learning model 104 for each divided image divided by the preprocessing program 102 . Specifically, the abnormality determination program 103 receives each divided image as an input, and performs determination processing by the learning model 104 to obtain an output of a determination result as to whether the image is normal or abnormal. The learning model 104 uses teacher data obtained by combining the image of the flange portion of the roof member with normal or abnormal information as a label, for example, a CNN (Convolutional Neural Network), which is an example of supervised learning of machine learning. It is a learning model generated by a network). In the case of CNN, the divided images divided from the extracted image by the preprocessing program 102 must be divided into images (square images) having the same number of vertical and horizontal pixels. However, it is not always necessary to directly divide the extracted image into square images, and it is possible to change the image size so that the divided image becomes a square within a range in which the accuracy of abnormality determination by the learning model 104 can be secured. good. Note that the learning algorithm for generating the learning model 104 is not limited to CNN, and may be generated by other learning algorithms. Further, in the present embodiment, as a result of abnormality determination processing using the learning model 104 from the image of the flange portion, the result of determining whether the shape of the flange portion is normal or abnormal and the degree of reliability of the determination result are obtained. It is assumed that the indicated reliability is obtained.

The display device 13 is a display device such as an LCD (Liquid Crystal Display) or an OELD (Organic Electro-Luminescent Display) that displays the results of determination by the abnormality determination program 103 .

The server device 20 is a server device that accumulates determination results and the like by the abnormality determination program 103 .

(Hardware configuration of anomaly detection device)
FIG. 2 is a diagram illustrating an example of the hardware configuration of the abnormality detection device according to the embodiment; The hardware configuration of the abnormality detection device 10 according to this embodiment will be described with reference to FIG.

As shown in FIG. 2, the abnormality detection device 10 includes a CPU (Central Processing Unit) 201, a RAM (Random Access Memory) 202, a ROM (Read Only Memory) 203, an auxiliary storage device 204, and a network I/F 205. , an input/output I/F 206 , a display output I/F 207 , and an imaging I/F 208 . These units are connected via a bus line 210 so as to enable data communication with each other.

The CPU 201 is an arithmetic device that controls the operation of the abnormality detection device 10 as a whole. A RAM 202 is a volatile storage device used as a work area for the CPU 201 . The ROM 203 is a non-volatile storage device that stores programs such as an IPL (Initial Program Loader) that the CPU 201 executes first. Note that the CPU 201, the RAM 202, and the ROM 203 may be configured as an SoC (System on a Chip) mounted on one substrate, for example.

Auxiliary storage device 204 stores image processing program 101, preprocessing program 102, abnormality determination program 103, learning model 104, and various data used in these programs. State Drive), eMMC (embedded Multi Media Card), or an auxiliary storage device such as a microSD card.

A network I/F 205 is an interface for connecting to a network or the like for data communication. In the example shown in FIG. 2 , the network I/F 205 is connected to the server device 20 and transmits the above-described determination results and the like of the abnormality detection device 10 to the server device 20 . Also, the network I/F 205 is an Ethernet (registered trademark) standard interface that enables communication using TCP (Transmission Control Protocol)/IP (Internet Protocol) protocols, for example. Note that the network I/F 205 is not limited to wired communication, and may be a wireless communication interface based on Wi-Fi (registered trademark) or the like.

The input/output I/F 206 is an interface such as GPIO (General Purpose Input/Output) for communicating with the PLC 11 . The input/output I/F 206 , for example, receives the carrier signal from the PLC 11 and transmits to the PLC 11 a signal indicating the result of determination by the abnormality determination program 103 (signal indicating normality or abnormality).

The display output I/F 207 is an interface such as HDMI (registered trademark) (High-Definition Multimedia Interface) for transmitting display data such as determination results by the abnormality determination program 103 to the display device 13 .

The imaging I/F 208 is an interface such as a USB (Universal Serial Bus) for receiving captured images from the imaging device 12 .

Note that the hardware configuration of the abnormality detection device 10 shown in FIG. 1 is an example, and is not limited to this configuration. For example, it may include a GPU (Graphics Processing Unit) that specializes in graphic processing.

(Functional block configuration of anomaly detection device)
FIG. 3 is a diagram illustrating an example of a functional block configuration of the anomaly detection device according to the embodiment; FIG. 4 is a diagram showing an example of an extracted image including a picked-up image of the roof member and a flange portion. FIG. 5 is a diagram showing an example of a normal image of the flange portion. FIG. 6 is a diagram showing an example of an abnormal image of the flange portion. FIG. 7 is a diagram for explaining a state in which the extracted image of the flange portion is divided. FIG. 8 is a diagram showing an example of a superimposed display of the determination result regarding the shape of the flange portion of the captured image. The functional block configuration and operation of the abnormality detection device 10 according to the present embodiment will be described with reference to FIGS. 3 to 8. FIG.

As shown in FIG. 3, the abnormality detection device 10 includes an acquisition unit 301, a process communication unit 302, an extraction unit 303, a preprocessing unit 304 (dividing unit), an abnormality determination unit 305 (determination unit), and a display It has a control unit 306 , a generation unit 307 and a storage unit 308 .

The acquisition unit 301 is a functional unit that acquires, via the imaging I/F 208 , a captured image including the flange portion of the roof member of the component 30 that is the vehicle body, captured by the imaging device 12 . Here, FIG. 4A shows a captured image IMG1 including the flange portion 31a of the roof member 31 of the component 30 captured by the imaging device 12 as an example of the captured image. In addition, when the acquisition unit 301 receives from the process communication unit 302 described later that the above-described transport signal regarding the component 30 has been received, the acquisition unit 301 starts acquisition of the captured image captured by the imaging device 12 . Note that the imaging operation by the imaging device 12 and the acquisition operation of the captured image by the acquisition unit 301 may be started at the timing when the carrier signal is received by the process communication unit 302 . The acquisition unit 301 outputs the acquired captured image to the extraction unit 303 . The acquisition unit 301 is realized, for example, by executing the image processing program 101 by the CPU 201 shown in FIG.

The process communication unit 302 is a functional unit that communicates via the input/output I/F 206 . For example, the process communication unit 302 receives a transport signal from the PLC 11 at the timing when the component 30 is transported to a predetermined position suitable for imaging by the imaging device 12 on the transport line 3, and notifies that the transport signal has been received. Output to acquisition unit 301 . In addition, the process communication unit 302 transmits a signal indicating the result of determination by the abnormality determination unit 305 (signal indicating normality or abnormality) to the PLC 11 . The process communication unit 302 is realized, for example, by executing a predetermined program by the CPU 201 shown in FIG.

The extraction unit 303 is a functional unit that extracts an image portion of a predetermined target object from the captured image acquired by the acquisition unit 301 . Specifically, the extraction unit 303 extracts the image of the flange portion of the roof member of the component 30 from the captured image acquired by the acquisition unit 301 . For example, in FIG. 4B, the extraction unit 303 extracts an image including the flange portion 31a of the roof member 31 of the component 30 from the captured image IMG1 shown in FIG. An example of extracting an extracted image EIMG1 is shown. An extracted image EIMG1 shown in FIG. 5 is an example of an extracted image including the normal-shaped flange portion 31a, which is extracted from the captured image by the extraction unit 303. As shown in FIG. On the other hand, an extracted image EIMG2 shown in FIG. 6 shows an example of an extracted image including the flange portion 31a having an abnormal shape extracted from the captured image by the extraction unit 303. In FIG. FIG. 6 shows an example in which the flange portion 31a shown in the extracted image EIMG2 includes three abnormal portions 31b as abnormal shape portions.

The extraction method by the extraction unit 303 is extraction processing of a predetermined region of the captured image, extraction processing based on a predetermined algorithm such as template matching, or YOLO using a predetermined learning model generated by supervised learning or the like. The method may be based on any existing extraction process. The extraction unit 303 outputs the extracted image of the flange portion 31 a to the preprocessing unit 304 . The extracting unit 303 is realized by executing the image processing program 101 by the CPU 201 shown in FIG. 2, for example.

The preprocessing unit 304 is a functional unit that executes predetermined preprocessing for abnormality determination processing by the abnormality determination unit 305 in the subsequent stage on the extracted image of the flange portion 31 a extracted by the extraction unit 303 . The preprocessing unit 304 divides the extracted image into a plurality of images as predetermined preprocessing, specifically, in order to reduce the load of abnormality determination processing by the abnormality determination unit 305 . In the example shown in FIG. 7, the preprocessing unit 304 divides the extraction image EIMG1 extracted by the extraction unit 303, for example, into five divided images SIMG. In addition, when the learning model 104 used by the abnormality determination unit 305 is based on CNN, as described above, the divided images obtained by dividing the extracted image are images having the same number of vertical and horizontal pixels (square images). must be split into However, it is not always necessary to directly divide the extracted image into square images, and it is possible to change the image size so that the divided image becomes a square within a range in which the accuracy of abnormality determination by the learning model 104 can be secured. good. The preprocessing unit 304 outputs the plurality of divided images to the abnormality determination unit 305 .

When the extraction image extracted by the extraction unit 303 is a high-resolution image with a large data size, the abnormality determination processing by the abnormality determination unit 305 is heavily loaded. By dividing the image into a plurality of divided images, the load of the abnormality determination process can be reduced. Further, in order to reduce the load of the abnormality determination processing by the abnormality determination unit 305, it is conceivable to reduce the data size by reducing the extracted image, but the resolution of the shape of the flange portion 31a included in the extracted image is reduced. This is not preferable because the accuracy of the abnormality determination process is lowered. In the example of FIG. 7, since the predetermined object is the horizontally long flange portion 31a, it is divided in the horizontal direction. Depending on the shape and size of the , it may be divided vertically or horizontally and vertically. In addition, it is not always necessary to divide the extracted image by the preprocessing unit 304, and depending on the data size and shape of the extracted image, the abnormality determination unit 305 directly executes abnormality determination processing on the extracted image. may be In this case, whether or not to divide the extracted image by the preprocessing unit 304 may be switched by setting, or may be automatically switched according to the data size or shape of the extracted image.

The preprocessing unit 304 is implemented by executing the preprocessing program 102 by the CPU 201 shown in FIG. 2, for example.

The abnormality determination unit 305 is a functional unit that executes abnormality determination processing using the learning model 104 stored in the storage unit 308 for each divided image divided by the preprocessing unit 304 . Specifically, the abnormality determination unit 305 receives each divided image as an input, and performs determination processing by the learning model 104 to output a determination result as to whether it is normal or abnormal (determination result indicating whether there is an abnormality). In this case, since the abnormality determination unit 305 performs abnormality determination processing for each divided image, it is possible to distinguish which divided image has an abnormality and output it. Further, the abnormality determination unit 305 outputs a determination result as to whether the shape of the flange portion 31a is normal or abnormal, and a degree of reliability indicating the degree of reliability of the determination result for each divided image. The abnormality determination unit 305 outputs the determination result and reliability for each divided image to the display control unit 306 and the server device 20 , and outputs the determination result for each divided image to the process communication unit 302 . Then, the process communication unit 302 transmits the determination result for each divided image to the PLC 11, and the PLC 11 performs predetermined processing on the transport line 3 when there is an abnormality in the determination result of any of the divided images. (line stop, etc.). The abnormality determination unit 305 is implemented by executing the abnormality determination program 103 by the CPU 201 shown in FIG. 2, for example.

The abnormality determination processing by the abnormality determination unit 305 is assumed to use the discrimination processing using the learning model 104, but is not limited to this, and may be performed using a predetermined algorithm such as template matching that does not use the learning model. It is also possible to execute the abnormality determination process based on the Even in this case, by dividing the extracted image into a plurality of divided images by the preprocessing unit 304, the load of the abnormality determination process can be reduced.

The display control unit 306 is a functional unit that causes the display device 13 to display the determination result received from the abnormality determination unit 305 . For example, the display control unit 306 superimposes and displays the determination result on the captured image acquired by the acquisition unit 301 . In addition to the determination result, the reliability of the determination result may be superimposed on the captured image and displayed. In the example shown in FIG. 8, the display control unit 306 indicates an abnormality in the abnormal portion of the flange portion 31a determined by the abnormality determination processing by the abnormality determination unit 305 in the captured image IMG2 acquired by the acquisition unit 301. "anomaly" as the determination result and superimposed display parts SPNa and SPNb including the reliability of the abnormality determination process are superimposed and displayed. It should be noted that the display control unit 306 may, for example, cause the display color of the superimposed display unit to be displayed differently depending on the numerical value of the reliability.

By displaying the determination result on the display device 13 in this way, the operator of the transfer line 3 can visually recognize the determination result. In addition, by superimposing the determination result on the abnormal portion of the flange portion 31a of the captured image and displaying it, the operator can easily confirm the abnormal portion of the shape of the flange portion 31a determined by the learning model 104. can be done. Further, by superimposing the reliability of the identification processing using the learning model 104 on the abnormal portion of the flange portion 31a of the captured image and displaying it, the reliability of the abnormality determination processing based on the learning model 104 can be confirmed. .

The display control unit 306 is realized, for example, by executing a program such as the abnormality determination program 103 by the CPU 201 shown in FIG.

The generation unit 307 generates the learning model 104 in advance by CNN, which is an example of supervised learning of machine learning, for example, using supervised data obtained by combining the image of the flange portion of the roof member with normal or abnormal information as a label. This is the function part to generate. The generation unit 307 causes the storage unit 308 to store the generated learning model 104 . The generation unit 307 is implemented by, for example, the CPU 201 shown in FIG. 2 executing a learning program that implements the above-described supervised learning.

The storage unit 308 is a functional unit that stores the image processing program 101, the preprocessing program 102, the abnormality determination program 103, the learning model 104, and various data used in these programs. As shown in FIG. 1 above, the image processing program 101, the preprocessing program 102, and the abnormality determination program 103 using the learning model 104 are separate programs. , may be configured as one program. The storage unit 308 is implemented by the auxiliary storage device 204 shown in FIG.

Note that each functional unit of the abnormality detection device 10 shown in FIG. 3 conceptually shows the function, and is not limited to such a configuration. For example, a plurality of functional units illustrated as independent functional units in FIG. 3 may be configured as one functional unit. On the other hand, the function of one functional unit in FIG. 3 may be divided into a plurality of functions to form a plurality of functional units.

(Flow of anomaly detection processing of the anomaly detection system)
FIG. 9 is a flowchart showing an example of the flow of anomaly detection processing of the anomaly detection system according to the embodiment. The flow of abnormality detection processing of the abnormality detection system 1 according to the present embodiment will be described with reference to FIG. 9 .

<Step S11>
First, the process communication unit 302 receives a transport signal from the PLC 11 indicating that the component 30 has been transported to a predetermined position on the transport line 3 at a timing suitable for imaging by the imaging device 12. If so (step S11: Yes), the abnormality detection device 10 returns from the standby mode to the normal operation mode, the process communication unit 302 outputs to the acquisition unit 301 that the transport signal has been received, and step S12 Move to If the component 30 has not been transported to the predetermined position (step S11: No), it waits until it is transported to the predetermined position. Here, the standby mode is, for example, an operation mode in a power saving state.

<Step S12>
Upon receiving the receipt of the carrier signal from the process communication unit 302 , the acquisition unit 301 starts acquiring the captured image captured by the imaging device 12 and outputs the acquired captured image to the extraction unit 303 . Then, the process proceeds to step S13.

<Step S13>
The extraction unit 303 extracts an image portion of the flange portion 31a of the roof member 31 of the component 30 as a predetermined object from the captured image acquired by the acquisition unit 301 . The extraction unit 303 outputs the extracted image of the flange portion 31 a to the preprocessing unit 304 . Then, the process proceeds to step S14.

<Step S14>
The preprocessing unit 304 divides the extraction image of the flange portion 31a extracted by the extraction unit 303 into a plurality of images as preprocessing for the abnormality determination processing by the abnormality determination unit 305 in the subsequent stage. The preprocessing unit 304 outputs the plurality of divided images to the abnormality determination unit 305 . Then, the process proceeds to step S15.

<Step S15>
The abnormality determination unit 305 executes abnormality determination processing using the learning model 104 stored in the storage unit 308 for each divided image divided by the preprocessing unit 304 . Specifically, the abnormality determination unit 305 receives each divided image as an input, performs determination processing by the learning model 104, and outputs a determination result as to whether the image is normal or abnormal. Further, the abnormality determination unit 305 outputs a determination result as to whether the shape of the flange portion 31a is normal or abnormal, and a degree of reliability indicating the degree of reliability of the determination result for each divided image. The abnormality determination unit 305 outputs the determination result and reliability for each divided image to the display control unit 306 and the server device 20 , and outputs the determination result for each divided image to the process communication unit 302 . Then, the process proceeds to step S16.

<Step S16>
The display control unit 306 superimposes and displays the determination result on the captured image acquired by the acquisition unit 301 . Then, the process proceeds to step S17.

<Step S17>
The process communication unit 302 transmits the determination result for each divided image to the PLC 11 . The PLC 11 executes a predetermined process (line stop, etc.) on the transport line 3 when any of the determination results of the divided images indicates an abnormality. Then, the abnormality detection device 10 shifts to the standby mode. Then, the abnormality detection process ends.

As described above, in the abnormality detection device 10 according to the present embodiment, the acquisition unit 301 captures an image of the flange portion 31a of the roof member 31 of the component 30 transported on the transport line 3, which is captured by the imaging device 12. is acquired, the preprocessing unit 304 divides the image portion of the flange portion 31a in the captured image acquired by the acquisition unit 301 into a plurality of images, and the abnormality determination unit 305 divides the plurality of images divided by the preprocessing unit 304 The divided image is input, and an abnormality determination process using the learning model 104 is performed to obtain a determination result indicating the presence or absence of an abnormality as an output. As a result, it is possible to accurately detect an abnormality in the shape of the flange portion 31a as a predetermined object. In addition, since the image portion of the flange portion 31a of the captured image is divided into a plurality of divided images, the load of abnormality determination processing by the abnormality determination unit 305 can be reduced.

Further, in the abnormality detection device 10, the extraction unit 303 extracts the image portion of the flange portion 31a from the captured image acquired by the acquisition unit 301 as an extracted image, and the preprocessing unit 304 extracts the image extracted by the extraction unit 303. It is assumed that the image is divided into a plurality of images. As a result, the abnormality determination process is performed after extracting the image portion including the flange portion 31a as the predetermined object, so that the accuracy of abnormality detection for the flange portion 31a can be improved.

In the abnormality detection device 10, the process communication unit 302 receives a transport signal indicating that the part 30 has been transported to a predetermined position on the transport line 3, and the acquisition unit 301 receives the transport signal from the process communication unit 302. Acquisition of a captured image captured by the imaging device 12 is started when the command is received. As a result, the processing load of the abnormality detection device 10 can be reduced since the captured image acquisition processing is started from the time the carrier signal is received and a series of processing is executed up to the abnormality determination processing.

In the abnormality detection device 10, when the process communication unit 302 receives the transport signal, the standby mode is resumed, and when the abnormality determination processing by the abnormality determination unit 305 is completed, the abnormality detection device 10 is switched to the standby mode. Thereby, a single board computer or the like can be adopted as the abnormality detection device 10, and the processing load can be reduced.

1 Abnormality Detection System 3 Transfer Line 10 Abnormality Detection Device 11 PLC
12 imaging device 13 display device 20 server device 30 part 31 roof member 31a flange 31b abnormal portion 101 image processing program 102 preprocessing program 103 abnormality determination program 104 learning model 301 acquisition unit 302 process communication unit 303 extraction unit 304 preprocessing unit 305 Abnormality determination unit 306 Display control unit 307 Generation unit 308 Storage unit

Claims (6)

an acquisition unit that acquires a captured image of a predetermined object in a component that is transported on a transport route, which is captured by an imaging device;
a division unit that divides an image portion of the object in the captured image acquired by the acquisition unit into a plurality of images;
a determining unit that receives the plurality of images divided by the dividing unit as input and obtains as an output a determination result indicating the presence or absence of an abnormality through abnormality determination processing using a predetermined learning model;
Anomaly detection device with further comprising an extraction unit that extracts an image portion of the object from the captured image acquired by the acquisition unit as an extracted image,
The anomaly detection device according to claim 1, wherein the dividing section divides the extracted image extracted by the extracting section into the plurality of images. further comprising a receiving unit that receives a signal indicating that the component has been transported to a predetermined position on the transport route;
The anomaly detection device according to claim 1 or 2, wherein the acquisition unit starts acquisition of the captured image captured by the imaging device when the signal is received by the reception unit. 4. The abnormality detection device according to claim 3, wherein when the signal is received by the receiving section, the standby mode is returned, and when the abnormality determination processing by the determination section is completed, the abnormality detection device is shifted to the standby mode. Claims 1 to 4, further comprising a generation unit that generates the learning model by supervised learning using teacher data that combines an image of a predetermined object in the part and a label indicating whether it is normal or abnormal. The abnormality detection device according to any one of 1. an acquisition step of acquiring a captured image of a predetermined object in a component transported on a transport route, which is imaged by an imaging device;
a dividing step of dividing the image portion of the object in the acquired captured image into a plurality of images;
a judgment step of obtaining, as an output, a judgment result indicating the presence or absence of an abnormality through an abnormality judgment process using a predetermined learning model using the plurality of divided images as an input;
An anomaly detection method comprising:

Images