An efficient industrial intelligent visual quality inspection method and device combined with a multi-modal network

Abstract

The application provides a kind of efficient industrial intelligent visual quality inspection method and device combined with multi-modal network, the method comprises: collecting tin paste spot image, forming quality inspection image;Quality inspection image is handled and named, and quality evaluation is carried out on tin paste block by image positioning segmentation, image detection identification, unqualified information reporting and other operations;Image file acquisition request sending, image file forwarding and the like.The quality inspection application of the present application combined with multi-modal network can avoid repeated transmission of the same quality inspection image, improve the transmission efficiency of quality inspection image;It can be applied to intelligent manufacturing, and can reduce the transmission overhead of image file.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of industrial internet, and in particular to an efficient industrial intelligent visual quality inspection method and device combining a multi-modal network. BACKGROUND

[0002] In the field of electronic circuit patch processing, tin paste detection is a key link in quality control. The quality inspection principle is to output the reflected light intensity of the detected object as a quantitative gray value using camera technology, compare it with the gray value of the standard image, analyze and determine the defects and classify them. With the development of artificial intelligence technology, using artificial intelligence technology can effectively improve the quality inspection efficiency of tin paste detection, and plays an important role in quality control, production process, operation and maintenance of products.

[0003] In the current artificial intelligence-based quality inspection method, the quality inspection terminal shoots high-definition videos and images of product quality, uploads them to the detection system deployed on the industrial cloud platform for evaluation, and alarms unqualified products.

[0004] The existing technology has the following defects: multiple business links in the factory will frequently call photos recording defect information, and through obtaining these defect information, the product quality defects are rechecked to better iterate and update, and the enterprise's product lifecycle management (PLM) system also needs to obtain these information to provide product whole-process traceability capability. The existing method repeatedly transmits the same content multiple times in the transmission process, and the transmission efficiency is low. SUMMARY

[0005] The purpose of the present application is to provide an efficient industrial intelligent visual quality inspection method and device combining a multi-modal network, which solves the problems of repeated transmission of the same quality inspection image and low transmission efficiency in the prior art.

[0006] To achieve the above purpose, the present application provides an efficient industrial intelligent visual quality inspection method combining a multi-modal network, which comprises the following steps:

[0007] Step 1: Build a quality inspection image transmission network composed of programmable switches, and deploy a quality inspection model on the data center server in the quality inspection image transmission network;

[0008] Step 2: Collect tin paste spot images by a quality inspection camera, form quality inspection images after pretreatment of the tin paste spot images, and upload the quality inspection images to the data center;

[0009] Step three, obtain the quality inspection image from the data center through the quality inspection model, segment to obtain each tin paste block image, evaluate the quality of the tin paste printing, judge whether there is an unqualified tin paste block in the detected tin paste block, and feed back the evaluation result to the data center; if an image containing an unqualified tin paste block is detected, the data center sends a message to the factory control center, warning that unqualified products have occurred, and the message content is the corresponding quality inspection image name; the factory control center forwards the alarm information to multiple related departments; each related department sends a request to the data center to obtain the corresponding quality inspection image, the request is sent in the form of an interest package, wherein the interest package contains the hash value of the name corresponding to the quality inspection image; if it is detected that the images of the tin paste blocks are all qualified, return to step two to collect new tin paste spot images;

[0010] Step four, after the programmable switch in the quality inspection image transmission network receives the interest package, the name hash value is parsed; the data center returns the corresponding image file in the form of a data package; after the programmable switch receives the data package, the name hash value in the data package is parsed, the data package is forwarded to the interest package port, and the data package is also forwarded to the storage module; after multiple hop forwarding, the data package returns to the process department that requested the quality inspection image file;

[0011] Step five, the process department receives the image and further analyzes and improves the process flow;

[0012] Step six, another related department sends a request to the data center to obtain the same quality inspection image file from the nearest programmable switch.

[0013] Further, in step one, the image transmission forwarding rule is implemented based on programmable switch forwarding, multiple programmable switches are controlled by a controller, the controller deploys a network mode suitable for the quality inspection application, that is, a content identification mode, to the programmable switch; each programmable switch has a storage module that stores the hash value of the quality inspection image name and the corresponding quality inspection image content.

[0014] Further, in step two, the quality inspection camera is deployed on the intelligent manufacturing production line and connected to the programmable switch, and finally forwarded to the data center server through multiple programmable switches.

[0015] Further, in step two, the name of the quality inspection image is production line name-product serial number-current timestamp, and the data is uploaded to the data center in the form of IP for archiving and evaluation.

[0016] Further, the step three is specifically: obtaining a quality inspection image from the data center through a quality inspection model, determining the solder paste block boundary based on pixel changes, further segmenting to obtain each solder paste block image, evaluating the quality of the solder paste printing, judging whether there is an unqualified solder paste block in the detected solder paste block, and feeding back the evaluation result to the data center by the quality inspection model; if an image containing an unqualified solder paste block is detected, the data center sends a message to the factory control center to alarm the occurrence of unqualified products, and the message content is the corresponding quality inspection image name; the factory control center forwards the alarm information to multiple related departments according to the production line name and product serial number; each related department sends a request to the data center to obtain the corresponding quality inspection image; the request is sent in the form of an interest package, wherein the interest package contains the hash value of the image corresponding to the name; if it is detected that the images of the solder paste blocks are all qualified, return to step two to collect new solder paste spot images.

[0017] Further, the step four includes the following sub-steps:

[0018] (4.1) The programmable switch in the quality inspection image transmission network receives the interest package and analyzes the name hash value; if the interest package is the first request, the interest package is forwarded to other programmable switches, and the response or final forwarding to the data center is realized by other programmable switches;

[0019] (4.2) After the data center receives the interest package, the hash value of the name in the interest package is analyzed, the name is calculated, and the image file to be sent is found according to the name; then, according to the segment size preset by the storage module, the number of segments to which the image file to be sent should be divided is calculated, the content of each segment of the image to be sent is formed and named; the hash value of the name of each segment of the image to be sent is calculated using a preset hash function; it is judged whether the hash value of the name of each segment of the image to be sent conflicts with the hash value of the name that has been stored, if a hash conflict occurs, the hash value of the name of each segment of the image is recalculated using a candidate hash function until no conflict occurs; the content is segmented and encapsulated, and the hash value of the name of each segment of the image content, the number of segments to which the corresponding image file should be divided, the current segment number, and the image content of the current segment are recorded respectively, thereby completing the quality inspection image data package packaging; the packaged data package is sent out by the data center;

[0020] (4.3) After the programmable switch receives the data package, the hash value of the name in the data package is analyzed, and the data package is forwarded to the interest package port, which is the port that has ever received the name identifier corresponding to the content in the data package; at the same time, the programmable switch forwards the data package to the storage module, and the storage module stores the hash value of the name in the data package, the current segment number, the total number of segments, and the image content; after multiple hop forwarding, the data package returns to the process department that requests the quality inspection image file.

[0021] Further, the step six is specifically: another relevant department sends a request to the data center to obtain the same quality inspection image file; when the interest package arrives at the programmable switch in the quality inspection image transmission network, the programmable switch judges that the same request has been responded to previously, and is directly forwarded to the storage module; the storage module receives the interest package, and then finds the corresponding quality inspection image content according to the named hash value in the interest package, and encapsulates the quality inspection image content into a data package, the data package containing the hash value of the quality inspection image name and the corresponding quality inspection image content; the storage module forwards the data package to the interest package port of the programmable switch, and the interest package port is a port that has received the name containing the content corresponding to the data package; thus, the other department obtains the image file from the nearest programmable switch.

[0022] To achieve the above object, the application further provides a high-efficiency industrial intelligent visual quality inspection device combined with a multi-modal network, comprising one or more processors for implementing the high-efficiency industrial intelligent visual quality inspection method combined with a multi-modal network.

[0023] To achieve the above object, the application further provides an electronic device comprising a memory and a processor, wherein the memory is coupled with the processor; the memory is used to store program data, and the processor is used to execute the program data to implement the high-efficiency industrial intelligent visual quality inspection method combined with a multi-modal network.

[0024] To achieve the above object, the application further provides a computer readable storage medium having a computer program stored thereon, wherein the program is executed by a processor to implement the high-efficiency industrial intelligent visual quality inspection method combined with a multi-modal network.

[0025] The application has the advantages that the quality inspection application combined with a multi-modal network can avoid repeated transmission of the same quality inspection image, and improve the quality inspection image transmission efficiency. The application can be applied to intelligent manufacturing, and can reduce the transmission cost of image files. BRIEF DESCRIPTION OF DRAWINGS

[0026] In order to more clearly illustrate the technical solutions in the embodiments of the application, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0027] Figure 1 A flowchart of a high-efficiency industrial intelligent visual quality inspection method combined with a multi-modal network provided by an embodiment of the application;

[0028] Figure 2A structural block diagram of an efficient industrial intelligent visual quality inspection method device combined with a multi-modal network is provided for an embodiment of the present application.

[0029] Figure 3 A schematic diagram of an electronic device is shown for an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0030] The technical solutions in the embodiments of the present application will be clearly and completely described with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of the present application.

[0031] It should be noted that the features in the following embodiments and implementation manners can be combined with each other without conflict.

[0032] Figure 1 A flowchart of an efficient industrial intelligent visual quality inspection method combined with a multi-modal network is provided for an embodiment of the present application. As shown in Figure 1 The efficient industrial intelligent visual quality inspection method combined with a multi-modal network in the embodiments of the present application can include the following steps:

[0033] Step one, deploying a quality inspection system architecture, including building a quality inspection image transmission network composed of programmable switches, deploying a quality inspection model on a data center server in the quality inspection image transmission network, and accessing a quality inspection terminal.

[0034] Specifically, the quality inspection system architecture is deployed: an industrial real-time quality inspection model is deployed on a smart industrial park data center server. The quality inspection image transmission network is composed of programmable switches, and the image transmission forwarding rule is realized based on programmable switch forwarding. A plurality of programmable switches are controlled by a controller, and the controller deploys a network mode suitable for the quality inspection application, i.e., a content identification mode, to the programmable switches. Each programmable switch has a storage module that stores the hash value of the quality inspection image name and the corresponding quality inspection image content. A quality inspection camera is deployed on a smart manufacturing production line and connected to the programmable switch. And through the forwarding of multiple programmable switches, it is finally forwarded to the data center server.

[0035] Step two, collecting quality inspection images.

[0036] (2.1) Original photo shooting: start the quality inspection camera on the production line, shoot the printed circuit board photo, and collect the solder paste spot image.

[0037] (2.2) Image preprocessing: normalize and grayscale the original photo data. Then, pass through a filter to reduce noise and form a quality inspection image.

[0038] (2.3) The quality inspection camera generates a quality inspection image for the current product, named as the production line name-product serial number-current timestamp, and uploads the factory data center archive in IP mode for evaluation.

[0039] Step three, quality evaluation of the tin paste block.

[0040] (3.1) Image positioning and segmentation: the quality inspection model obtains the quality inspection image from the data center, determines the tin paste block boundary based on pixel changes, further segments to obtain each tin paste block image, and obtains the height, volume, area, shape and position of the printed tin paste and other key parameters.

[0041] (3.2) Image detection and identification: based on the above key parameters, the quality of tin paste printing is evaluated by comparing with the preset standard parameters, and it is judged whether there is unqualified tin paste block in the detected tin paste block.

[0042] (3.3) Quality inspection alarm; the quality inspection model feeds back the evaluation result to the data center. For the image containing unqualified tin paste block detected, the data center sends a message to the factory control center, alarming the occurrence of unqualified product, and the message content is the corresponding quality inspection image name.

[0043] (3.4) The factory control center forwards the alarm information to multiple related departments according to the production line name and product serial number. Each related department sends a request to the data center to obtain the corresponding quality inspection image. The request is sent in the form of interest package. The interest package contains the hash value of the image name.

[0044] (3.5) If the images of the tin paste block are all qualified, return to step two to collect new tin paste spot images.

[0045] Step four, request and acquisition of quality inspection image.

[0046] (4.1) The programmable switch in the quality inspection image transmission network receives the interest package, analyzes the hash value of the name, and first checks whether the interest package has been received before. At the first request, since the programmable switch has not received the same interest package, the interest package is forwarded to other programmable switches, which respond or finally forward to the data center.

[0047] (4.2) The factory data center returns the corresponding image file in the form of data package.

[0048] Specifically, after the data center receives the interest package, the hash value of the name in the interest package is analyzed, and the name is calculated. The data center finds the image file to be sent according to the name;

[0049] The data center calculates the number of segments that the image file to be sent should be divided into according to the segment size preset by the storage module;

[0050] According to the number of segments, the image file to be sent is divided into segments in sequence to form each segment of image content to be sent;

[0051] Each segment of image content to be sent is named, and the naming method is to add a segment number to the name of the image file to be sent;

[0052] The hash value of the name of each segment of image content to be sent is calculated using a preset hash function;

[0053] It is judged whether the hash value of the name of each segment of image content to be sent conflicts with the hash value of the name that has been stored. If a hash conflict occurs, the hash value of the name of each segment of image content is recalculated using a candidate hash function until no conflict occurs;

[0054] The content segments are encapsulated, and the hash value of the name of each segment of image content, the number of segments that the corresponding image file should be divided into, the current segment number, and the image content of the current segment are recorded respectively, thereby completing the packaging of the quality inspection image data packet;

[0055] The data center sends the encapsulated data packet.

[0056] (4.3) After the programmable switch receives the data packet, it parses the hash value of the name in the data packet, forwards the data packet to the interest packet port, which is the port that has received the name identification corresponding to the content in the data packet. At the same time, the programmable switch forwards the data packet to the storage module, which stores the hash value of the name in the data packet, the current segment number, the total number of segments, and the image content. After multiple hop forwarding, the data packet returns to the process department that requested the image file.

[0057] Step five, improvement and control of quality process: the process department receives the image for further analysis and improvement of the process flow.

[0058] Step six, repeated acquisition of the same quality inspection image.

[0059] Specifically, another related department sends a request to the data center to obtain the same quality inspection image file. When the interest packet arrives at the programmable switch in the quality inspection image transmission network, the programmable switch determines that the same request has been responded to previously, and thus directly forwards to the storage module. The storage module receives the interest packet, and then finds the corresponding quality inspection image content according to the named hash value in the interest packet, and encapsulates the quality inspection image content into a data packet, the data packet containing the hash value of the name of the quality inspection image and the quality inspection image content corresponding thereto. The storage module forwards the data packet to the interest packet port of the programmable switch, which is the port that has once received the content containing the name corresponding to the data packet. In this way, the other department obtains the image file from the nearby programmable switch.

[0060] Corresponding to the foregoing embodiment of the high-efficiency industrial intelligent visual quality inspection method combined with a multi-modal network, the present application also provides an embodiment of a high-efficiency industrial intelligent visual quality inspection device combined with a multi-modal network.

[0061] Referring to Figure 2 , the embodiment of the present application provides a high-efficiency industrial intelligent visual quality inspection device combined with a multi-modal network, which comprises one or more processors for implementing the high-efficiency industrial intelligent visual quality inspection method combined with a multi-modal network in the foregoing embodiments.

[0062] The embodiment of the high-efficiency industrial intelligent visual quality inspection device combined with a multi-modal network of the present application can be applied to any device with data processing capability, which can be a device or apparatus such as a computer. The device embodiment can be realized by software, or by hardware or a combination of software and hardware. Taking software realization as an example, as a logically meaningful device, it is formed by reading the corresponding computer program instructions in the non-volatile memory into the memory for running by the processor of the device with data processing capability where it is located. From the hardware level, as shown in Figure 2 , it is a hardware structure diagram of the device with data processing capability where the high-efficiency industrial intelligent visual quality inspection device combined with a multi-modal network of the present application is located. In addition to the processor, the memory, the network interface, and the non-volatile memory shown in Figure 2 , the device with data processing capability where the device in the embodiment is located usually includes other hardware according to the actual functions of the device with data processing capability, and details are not described herein.

[0063] The implementation process of the functions and roles of each unit in the above device is specifically described in the implementation process of the corresponding steps in the above method, and is not described herein.

[0064] For the device embodiment, since it basically corresponds to the method embodiment, the relevant part can be seen from the part of the method embodiment. The device embodiment described above is only illustrative, wherein the units described as separate components can or can not be physically separated, and the components displayed as units can or can not be physical units, i.e., can be located in one place or distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the present application according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0065] Corresponding to the foregoing embodiments of the efficient industrial intelligent visual quality inspection method combining the multi-modal network, the embodiments of the present application also provide an electronic device, comprising: one or more processors; a memory for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the efficient industrial intelligent visual quality inspection method combining the multi-modal network as described above. As Figure 3 As shown in the figure, the efficient industrial intelligent visual quality inspection method combining the multi-modal network provided by the embodiments of the present application is a hardware structure diagram of any data processing capable device. In addition to the Figure 3 In addition to the processor, memory, DMA controller, disk, and non-volatile memory shown in the figure, any data processing capable device in which the device is located in the embodiment usually includes other hardware according to the actual function of the data processing capable device, which will not be described here.

[0066] Corresponding to the foregoing embodiments of the efficient industrial intelligent visual quality inspection method combining the multi-modal network, the embodiments of the present application also provide a computer readable storage medium having a program stored thereon, which is executed by a processor to implement the efficient industrial intelligent visual quality inspection method combining the multi-modal network in the above embodiments.

[0067] The computer readable storage medium can be an internal storage unit of any data processing capable device, such as a hard disk or a memory. The computer readable storage medium can also be any data processing capable device, such as a plug-in hard disk, a smart media card (SMC), an SD card, a flash card, etc. Further, the computer readable storage medium can include both the internal storage unit of any data processing capable device and an external storage device. The computer readable storage medium is used to store the computer program and other programs and data required by the data processing capable device, and can also be used to temporarily store data that has been output or will be output.

[0068] The above merely describes preferred embodiments of the present application, but is not used to limit the present application, and any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0069] The above embodiments are merely used to illustrate the design ideas and characteristics of the present application, and the purpose is to enable those skilled in the art to understand the present application and implement it, and the protection scope of the present application is not limited to the above embodiments. Therefore, any equivalent change or modification made according to the disclosed principles and design ideas of the present application shall be included in the protection scope of the present application.

Claims

1. An efficient industrial intelligent visual quality inspection method combined with a multi-modal network, characterized in that, The method comprises the following steps: Step one, build a quality inspection image transmission network composed of programmable switches, and deploy a quality inspection model on the data center server in the quality inspection image transmission network; the image transmission forwarding rule is realized based on programmable switch forwarding, a plurality of programmable switches are controlled by a controller, the controller deploys a network mode suitable for the quality inspection application, i.e. a content identification mode, to the programmable switches; each programmable switch has a storage module; Step two, collect the solder paste spot image through the quality inspection camera, form the quality inspection image after preprocessing the solder paste spot image, and upload the quality inspection image to the data center; Step three, obtain the quality inspection image from the data center through the quality inspection model, segment to obtain each solder paste block image, evaluate the quality of the solder paste printing, judge whether there is an unqualified solder paste block in the detected solder paste block, and feed back the evaluation result to the data center; if an image containing an unqualified solder paste block is detected, the data center sends a message to the factory control center to alarm the presence of unqualified products, and the message content is the corresponding quality inspection image name; the factory control center forwards the alarm information to a plurality of related departments; each related department sends a request to the data center to obtain the corresponding quality inspection image, and the request is sent in the form of an interest package, wherein the interest package contains the hash value of the corresponding name of the quality inspection image; if the images of the solder paste blocks are all qualified, return to step two to collect new solder paste spot images; Step four, after the programmable switch in the quality inspection image transmission network receives the interest package, the name hash value is analyzed; if the interest package is the first request, the interest package is forwarded to other programmable switches, and the other programmable switches respond or finally forward to the data center; the data center returns the corresponding image file in the form of a data package; the programmable switch receives the data package, analyzes the hash value of the name in the data package, forwards the data package to the interest package port, and the interest package port is the port that has received the name identification corresponding to the content in the data package; at the same time, the data package is forwarded to the storage module; after multi-hop forwarding, the data package returns to the process department requesting the quality inspection image file; Step five, the process department receives the image to further analyze and improve the process flow; Step six, another related department sends a request to the data center to obtain the same quality inspection image file from the nearest programmable switch.

2. The method of claim 1, wherein, In step one, the storage module stores the hash value of the quality inspection image name and the corresponding quality inspection image content.

3. The method of claim 1, wherein, In step two, the quality inspection camera is deployed on the intelligent manufacturing production line, connected with the programmable switch, and finally forwarded to the data center server through a plurality of programmable switches.

4. The method of claim 1, wherein, In step two, the name of the quality inspection image is production line name-product serial number-current timestamp, and the quality inspection image is uploaded to the data center in the form of IP for archiving and evaluation.

5. The method of claim 1, wherein, The third step is specifically: obtaining a quality inspection image from the data center through a quality inspection model, determining the solder paste block boundary based on pixel changes, further segmenting to obtain each solder paste block image, evaluating the quality of the solder paste printing, judging whether there is an unqualified solder paste block in the detected solder paste block, and feeding back the evaluation result to the data center by the quality inspection model; if an image containing an unqualified solder paste block is detected, the data center sends a message to the factory control center to alarm the presence of unqualified products, and the message content is the corresponding quality inspection image name; the factory control center forwards the alarm information to multiple related departments according to the production line name and product serial number; each related department sends a request to the data center to obtain the corresponding quality inspection image; the request is sent in the form of an interest package, wherein the interest package contains the hash value of the image corresponding to the name; if the images of the solder paste blocks are all qualified, return to step two to collect new solder paste spot images.

6. The method of claim 1, wherein, The fourth step includes the following sub-steps: (4.1) After the programmable switch in the quality inspection image transmission network receives the interest package, the name hash value is analyzed; if the interest package is the first request, the interest package is forwarded to other programmable switches, and the response or final forwarding to the data center is performed by other programmable switches; (4.2) After the data center receives the interest package, the hash value of the name in the interest package is analyzed, the name is calculated, and the image file to be sent is searched according to the name; then, according to the segment size preset by the storage module, the number of segments to be divided for the image file to be sent is calculated, the content of each segment of the image to be sent is formed and named; the hash value of the name of each segment of the image to be sent is calculated using a preset hash function; it is judged whether the hash value of the name of each segment of the image to be sent conflicts with the hash value of the name that has been stored, if a hash conflict occurs, the hash value of the name of each segment of the image is recalculated using a candidate hash function until no conflict occurs; the content is segmented and encapsulated, and the hash value of the name of each segment of the image content, the number of segments to be divided for the corresponding image file, the current segment number, and the image content of the current segment are recorded, thereby completing the quality inspection image data package packaging; the packaged data package is sent out by the data center; (4.3) After the programmable switch receives the data package, the hash value of the name in the data package is analyzed, and the data package is forwarded to the interest package port, which is the port that has received the name identifier corresponding to the content in the data package; at the same time, the programmable switch forwards the data package to the storage module, and the storage module stores the hash value of the name in the data package, the current segment number, the total number of segments, and the image content; after multiple hop forwarding, the data package returns to the process department that requests the quality inspection image file.

7. The method of claim 1, wherein, The step six is specifically: another relevant department sends a request to the data center to obtain the same quality inspection image file; when the interest package arrives at the programmable switch in the quality inspection image transmission network, the programmable switch judges that the same request has been responded to previously, and is directly forwarded to the storage module; when the storage module receives the interest package, the corresponding quality inspection image content is searched according to the named hash value in the interest package, and the quality inspection image content is encapsulated into a data package, the data package containing the hash value of the quality inspection image name and the corresponding quality inspection image content; the storage module forwards the data package to the interest package port of the programmable switch, and the interest package port is the port that has received the name containing the content corresponding to the data package; thus, the other department obtains the image file from the nearest programmable switch.

8. An efficient industrial intelligent visual quality inspection device combined with a multi-modal network, characterized in that, The one or more processors are configured to implement the efficient industrial intelligent visual quality inspection method combined with a multi-modal network according to any one of claims 1-7.

9. An electronic device comprising a memory and a processor, characterized in that The memory is coupled to the processor, and is configured to store program data, and the processor is configured to execute the program data to implement the efficient industrial intelligent visual quality inspection method combined with a multi-modal network according to any one of claims 1-7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the efficient industrial intelligent visual quality inspection method combined with a multi-modal network according to any one of claims 1-7.

Images