Industrial visual defect detection system and method
By acquiring images in real time on industrial production lines and using ResNet, YOLO, and Mask R-CNN models for multidimensional defect detection, quality inspection reports are generated. This solves the problems of low efficiency and insufficient accuracy in traditional detection methods, and achieves efficient and accurate defect identification and quality control of industrial products.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- HANGZHOU BROADLINK ELECTRONICS TECH
- Filing Date
- 2025-12-12
- Publication Date
- 2026-05-15
AI Technical Summary
Traditional industrial product defect detection relies on manual visual inspection, which is inefficient and easily affected by subjective factors. Existing automated detection solutions have poor versatility and limited recognition accuracy, making it difficult to effectively distinguish complex defects.
An image acquisition module is used to acquire images of industrial products in real time. Multidimensional defect detection is performed using trained ResNet, YOLO, and Mask R-CNN network models. The output control module generates quality inspection reports based on defect type, location, and shape, enabling early warning processing at different levels.
It enables automated multi-dimensional defect identification of industrial products on the production line, improves the reliability of defect detection and the level of quality control, and solves the problems of low efficiency and insufficient accuracy in traditional detection methods.
Smart Images
Figure CN122048780A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of industrial production technology, and in particular to an industrial visual defect detection system and method. Background Technology
[0002] Traditional product defect detection relies heavily on manual visual inspection, which is inefficient and susceptible to subjective factors, resulting in low accuracy and failing to meet the demands of modern production for high-efficiency and high-precision quality control. Existing automated inspection solutions often use cameras and preset rules for detection, but this approach has poor versatility, limited recognition accuracy, and is ineffective at handling complex defects (such as "dark lines" on injection molded parts), making it difficult to effectively distinguish between various complex defects.
[0003] Currently, no effective solution has been proposed for improving the reliability of defect detection in industrial products using relevant technologies. Summary of the Invention
[0004] This application provides an industrial visual defect detection system and method to at least address the problem of how to improve the reliability of defect detection in industrial products in related technologies.
[0005] In a first aspect, embodiments of this application provide an industrial visual defect detection system, the system comprising an image acquisition module, a detection inference module, and an output control module; The image acquisition module is used to acquire images of industrial products in real time on the production line; The detection reasoning module is used to perform multi-dimensional defect detection based on the industrial product image using a trained defect detection model to obtain multi-dimensional defect information of the industrial product, wherein the multi-dimensional defect information includes defect type information, defect location information, and defect shape information. The output control module is used to perform different levels of early warning processing based on the multidimensional defect information and generate corresponding quality inspection reports.
[0006] In some embodiments, the detection inference module is used to perform multi-dimensional defect detection based on the industrial product image using a trained first defect detection model to obtain defect type information of the industrial product. Based on the image of the industrial product, multidimensional defect detection is performed using a trained second defect detection model to obtain the defect location information of the industrial product. Based on the image of the industrial product, multidimensional defect detection is performed using a trained third defect detection model to obtain the defect shape information of the industrial product.
[0007] In some embodiments, the first defect detection model is a ResNet network model, the second defect detection model is a YOLO network model, and the third defect detection model is a Mask R-CNN network model.
[0008] In some embodiments, the system includes a neural network training module; The neural network training module is used to train the first defect detection model, the second defect detection model, and the third defect detection model respectively based on the constructed training dataset, so as to obtain the trained defect detection model.
[0009] In some embodiments, the system includes a data annotation module; The data annotation module is used to perform multi-type label annotation on the collected industrial product images to construct a training dataset, wherein the multi-type label annotation includes defect type annotation, defect location annotation, and defect shape annotation.
[0010] In some embodiments, the output control module is configured to determine the level of a first warning message for the industrial product on the assembly line based on the defect type information; determine the level of a second warning message for the industrial product on the assembly line based on the defect location information; and determine the level of a third warning message for the industrial product on the assembly line based on the defect shape information; wherein the warning message levels include critical, severe, moderate, and minor levels. By combining the first warning information, the second warning information, and the level of the second warning information, different levels of warning processing are triggered, and corresponding quality inspection reports are generated.
[0011] In some embodiments, the output control module is used to integrate the first warning information, the second warning information, and the level of the second warning information to trigger different levels of warning processing: If any of the first warning message, the second warning message, or the third warning message contains a critical warning message, the production line operation shall be stopped immediately and emergency procedures shall be initiated. If any of the first warning message, the second warning message, or the third warning message is of a severe level, the production line shall be suspended for rectification and shall be put back into operation after the rectification is completed. If only a general or minor level warning message exists among the first warning message, the second warning message, and the third warning message, then there is no need to stop the operation of the production line and optimize the automatic tracking and recording.
[0012] In some embodiments, the system includes a model deployment module; The model deployment module is used to compress and optimize the trained defect detection model through model quantization and model pruning, so as to deploy it in edge devices on the production line.
[0013] In some embodiments, the system includes a data preprocessing module; The data preprocessing module is used to preprocess the industrial product images acquired in real time by the image acquisition module on the production line to obtain preprocessed industrial product images. The preprocessing includes noise reduction, brightness normalization, image enhancement, and size cropping.
[0014] Secondly, embodiments of this application provide an industrial visual defect detection method, the execution of which is based on the system described in the first aspect above, the method comprising: Real-time image acquisition of industrial products on the production line; Based on the image of the industrial product, multidimensional defect detection is performed using a trained defect detection model to obtain multidimensional defect information of the industrial product, wherein the multidimensional defect information includes defect type information, defect location information, and defect shape information. Based on the multidimensional defect information, different levels of early warning processing are performed, and corresponding quality inspection reports are generated.
[0015] Compared to related technologies, this application provides an industrial visual defect detection system and method. The system includes an image acquisition module for real-time acquisition of industrial product images on a production line; a detection inference module for performing multi-dimensional defect detection based on the industrial product images using a trained defect detection model to obtain multi-dimensional defect information, including defect type, defect location, and defect shape information; and an output control module for executing different levels of early warning processing based on the multi-dimensional defect information and generating corresponding quality inspection reports. This system achieves multi-dimensional defect detection of industrial products on the production line, automating the identification of various complex defects based on defect type, location, and shape to execute different levels of early warning processing, thereby improving the precision of quality control and solving the problem of how to improve the reliability of defect detection in industrial products. Attached Figure Description
[0016] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a structural block diagram of an industrial visual defect detection system according to an embodiment of this application; Figure 2 This is a flowchart of the steps of an industrial visual defect detection method according to an embodiment of this application; Figure 3 This is a schematic diagram of the internal structure of an electronic device according to an embodiment of this application. Detailed Implementation
[0017] To make the objectives, technical solutions, and advantages of this application clearer, the application is described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments provided in this application without inventive effort are within the scope of protection of this application.
[0018] Obviously, the accompanying drawings described below are merely some examples or embodiments of this application. Those skilled in the art can apply this application to other similar scenarios based on these drawings without any inventive effort. Furthermore, it is understood that although the efforts made in this development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this application, any changes to design, manufacturing, or production based on the technical content disclosed in this application are merely conventional technical means and should not be construed as insufficient disclosure of the content of this application.
[0019] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment that is mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this application may be combined with other embodiments without conflict.
[0020] Unless otherwise defined, the technical or scientific terms used in this application shall have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms “a,” “an,” “an,” “the,” and similar words used in this application do not indicate quantity limitation and may indicate singular or plural. The terms “comprising,” “including,” “having,” and any variations thereof used in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units not listed, or may include other steps or units inherent to these processes, methods, products, or devices. The terms “connected,” “linked,” “coupled,” and similar words used in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. “Multiple” used in this application refers to two or more. “And / or” describes the relationship between related objects, indicating that three relationships may exist; for example, “A and / or B” can represent: A alone, A and B simultaneously, and B alone. The character " / " generally indicates that the preceding and following objects are in an "or" relationship. The terms "first," "second," and "third" used in this application are merely to distinguish similar objects and do not represent a specific ordering of the objects.
[0021] This application provides an industrial visual defect detection system. Figure 1 This is a structural block diagram of an industrial visual defect detection system according to an embodiment of this application, such as... Figure 1 As shown, the system includes an image acquisition module, a detection and inference module, and an output control module; Image acquisition module, used to acquire real-time images of industrial products on the production line; It should be noted that the image acquisition module is equipped with an industrial-grade high-resolution camera (such as a CCD or CMOS camera) and a light source system (such as ring light, backlight, bar light, etc., selected according to the object being inspected) to acquire product images in real time on the production line.
[0022] The detection and reasoning module is used to perform multi-dimensional defect detection based on the image of the industrial product and through a trained defect detection model to obtain multi-dimensional defect information of the industrial product. The multi-dimensional defect information includes defect type information, defect location information and defect shape information. Specifically, the detection reasoning module is used to perform multi-dimensional defect detection based on the industrial product image using a trained first defect detection model to obtain the defect type information of the industrial product. It should be noted that the first defect detection model is a ResNet network model. ResNet has strong object classification capabilities, showing high classification accuracy for cracks, deformations, and missing corners in automotive parts; solder joint defects and component misalignment on PCB boards; and surface scratches, pores, and cracks in metal products. This provides a basis for subsequent early warning processing at different levels, improving the precision of quality control. Furthermore, it has low computational resource requirements, making it suitable for deployment in edge devices on the shop floor. For example, using ResNet-18 to build a defect detection model for automotive parts can achieve a response time of <10ms and a defect recognition rate of 99.2%.
[0023] Based on the images of industrial products, multidimensional defect detection is performed using a trained second defect detection model to obtain the defect location information of the industrial products. It should be noted that the second defect detection model is a YOLO network model. YOLO has strong target localization capabilities and can quickly identify the location information of the target, providing a basis for subsequent early warning processing at different levels and improving the precision of quality control; it is especially suitable for high-speed production line scenarios that require a detection speed of ≥30 FPS.
[0024] Based on images of industrial products, a trained third-party defect detection model is used to perform multi-dimensional defect detection, obtaining the defect shape information of the industrial products. It should be noted that the third defect detection model is the Mask R-CNN network model. Mask R-CNN has strong target segmentation capabilities and can accurately extract the contour information of the target, providing a basis for subsequent early warning processing at different levels and improving the precision of quality control.
[0025] The output control module is used to perform different levels of early warning processing based on multi-dimensional defect information and generate corresponding quality inspection reports.
[0026] Specifically, the output control module is used to determine the level of the first warning information for industrial products on the production line based on defect type information; the level of the second warning information based on defect location information; and the level of the third warning information based on defect shape information. The warning information levels include critical, severe, moderate, and minor levels. By combining the levels of the first, second, and third early warning information, different levels of early warning processing are triggered, and corresponding quality inspection reports are generated.
[0027] It should be noted that Table 1 is an example table showing the relationship between defect type information and the first warning information level, Table 2 is an example table showing the relationship between defect location information and the second warning information level, and Table 3 is an example table showing the relationship between defect shape information and the third warning information level.
[0028] Table 1
[0029] Table 2
[0030] Table 3
[0031] Preferably, the output control module is used to integrate the levels of the first warning information, the second warning information, and the third warning information to trigger different levels of warning processing: If any of the first, second, or third warning messages is of a critical level, the production line shall be stopped immediately and emergency procedures shall be initiated. If any of the first, second, or third warning messages is of a severe level, the production line will be suspended for rectification. It will be put back into operation after the rectification is completed. If only a general or minor level warning exists among the first, second, and third warning messages, there is no need to stop the production line operation and optimize the automatic tracking and recording.
[0032] It should be noted that, based on identifying the type, location, and shape of defects in industrial products on the production line, the levels of multi-level early warning information are determined to execute different levels of early warning processing. This achieves refined production line product quality control and significantly improves the reliability of defect detection in industrial products. Furthermore, the generated quality inspection report records all inspection results, including defect type, location, shape, confidence level, timestamp, early warning information, and early warning processing, and is traceable, providing data support for production process improvement.
[0033] The system provided in this application embodiment realizes multi-dimensional defect detection of industrial products on the production line. It automatically identifies various complex defects based on defect type, location, and shape, and performs early warning processing at different levels to improve the precision of quality control and solve the problem of how to improve the reliability of defect detection of industrial products.
[0034] In some embodiments, the system includes a neural network training module; The neural network training module is used to train the first defect detection model, the second defect detection model, and the third defect detection model respectively based on the constructed training dataset, so as to obtain the trained defect detection model.
[0035] It should be noted that the defect detection model can be built using deep learning framework tools (such as TensorFlow, PyTorch). Specifically, three algorithm network architectures, ResNet, YOLO, and Mask R-CNN, are selected respectively, and ResNet network models, YOLO network models, and Mask R-CNN network models are built in the above deep learning framework tools. Then, the built models are trained to obtain the first defect detection model, the second defect detection model, and the third defect detection model.
[0036] In some embodiments, the system includes a data annotation module; The data annotation module is used to perform multi-type label annotation on the collected industrial product images to build a training dataset. The multi-type label annotation includes defect type annotation, defect location annotation, and defect shape annotation.
[0037] It should be noted that the data annotation module provides a user-friendly graphical interface tool that supports the annotation of defect areas in images using rectangular boxes, polygons, etc., and specifies a specific type label (such as "scratches", "cold welds", "damage") and location (pixel coordinate information) for each defect.
[0038] In some embodiments, the system includes a model deployment module; The model deployment module is used to compress and optimize the trained defect detection model through model quantization and model pruning, so that it can be deployed in edge devices on the production line.
[0039] It should be noted that the trained neural network model is compressed, optimized, and deployed to edge computing devices to meet the real-time requirements of the production line. Model quantization and pruning techniques are used to reduce model size and inference time.
[0040] In some embodiments, the system includes a data preprocessing module; The data preprocessing module is used to preprocess the industrial product images acquired in real time by the image acquisition module on the production line to obtain preprocessed industrial product images. The preprocessing includes noise reduction, brightness normalization, image enhancement, and size cropping.
[0041] It should be noted that this data preprocessing module performs noise reduction, brightness normalization, image enhancement, and size cropping on the acquired raw images to ensure image quality and feature consistency. Furthermore, the image data collected for training the aforementioned network model can also be preprocessed using this data preprocessing module before annotation.
[0042] It should be further noted that the above modules can be functional modules or program modules, and can be implemented through software or hardware. For modules implemented through hardware, the above modules can reside in the same processor; or the above modules can be located in different processors in any combination.
[0043] This application provides an industrial visual defect detection method. Figure 2 This is a flowchart of the steps of the industrial visual defect detection method according to an embodiment of this application, as follows: Figure 2 As shown, the method includes the following steps: Step S202: Real-time acquisition of industrial product images on the production line; Step S204: Based on the industrial product image, perform multi-dimensional defect detection using a trained defect detection model to obtain multi-dimensional defect information of the industrial product. The multi-dimensional defect information includes defect type information, defect location information, and defect shape information. Step S206: Perform different levels of early warning processing based on multidimensional defect information and generate corresponding quality inspection reports.
[0044] It should be noted that steps S202 to S206 above are the main steps in industrial product inspection on the production line. If model training, deployment, and updates are included, the workflow is as follows: S1. Data Acquisition and Preprocessing: Industrial cameras acquire product images on the production line in real time and optimize them through an image preprocessing module.
[0045] S2. Training data preparation: Manually annotate a large number of product images offline or with the aid of online methods to distinguish between normal and abnormal samples, and perform fine classification and annotation of defects in abnormal samples.
[0046] S3. Neural Network Model Training: Input the labeled data into the neural network training module, select the appropriate network structure and parameters to train the model until the model performance reaches the preset index.
[0047] S4. Model Deployment and Optimization: Deploy the trained model to the production line testing equipment and optimize the model to meet real-time testing requirements.
[0048] S5. Real-time detection and feedback: Product images on the production line are used for real-time inference through deployed models, and the system outputs the defect type, location, shape and confidence level.
[0049] S6. Decision-making and execution: Based on the detection results, the system automatically triggers corresponding multi-level early warning processing.
[0050] S7. Iterative Optimization and Maintenance: Regularly evaluate the model, perform incremental training or model reconstruction based on new defective data or performance requirements, and continuously optimize system performance.
[0051] The method provided in this application embodiment realizes multi-dimensional defect detection of industrial products on the production line. It automatically identifies various complex defects based on defect type, location, and shape, so as to perform different levels of early warning processing, improve the precision of quality control, and solve the problem of how to improve the reliability of defect detection of industrial products.
[0052] It should be noted that the steps shown in the above process or in the flowchart of the accompanying figures can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.
[0053] This embodiment provides an electronic device including a memory and a processor. The memory stores a computer program, and the processor is configured to run the computer program to perform the steps in any of the above method embodiments.
[0054] Optionally, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor and the input / output device is connected to the processor.
[0055] Optionally, the electronic device may further include a processor, memory, network interface, display screen, and input device connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements an industrial visual defect detection method. The display screen may be a liquid crystal display (LCD) or an e-ink display. The input device may be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the device's casing, or an external keyboard, touchpad, or mouse.
[0056] It should be noted that the specific examples in this embodiment can refer to the examples described in the above embodiments and optional implementations, and will not be repeated here.
[0057] Furthermore, in conjunction with the industrial visual defect detection methods in the above embodiments, this application embodiment can provide a storage medium for implementation. This storage medium stores a computer program; when executed by a processor, the computer program implements any of the industrial visual defect detection methods in the above embodiments.
[0058] In one embodiment, Figure 3 This is a schematic diagram of the internal structure of an electronic device according to an embodiment of this application, such as... Figure 3 As shown, an electronic device is provided, which can be a server, and its internal structure diagram can be as follows. Figure 3 As shown, the electronic device includes a processor, a network interface, internal memory, and non-volatile memory connected via an internal bus. The non-volatile memory stores an operating system, computer programs, and a database. The processor provides computing and control capabilities, the network interface communicates with external terminals via a network, the internal memory provides an environment for the operation of the operating system and computer programs, the computer programs are executed by the processor to implement an industrial visual defect detection method, and the database stores data.
[0059] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the electronic device to which the present application is applied. A specific electronic device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.
[0060] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.
[0061] Those skilled in the art should understand that the technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments have been described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0062] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.
Claims
1. An industrial visual defect detection system, characterized in that, The system includes an image acquisition module, a detection and inference module, and an output control module; The image acquisition module is used to acquire images of industrial products in real time on the production line; The detection reasoning module is used to perform multi-dimensional defect detection based on the industrial product image using a trained defect detection model to obtain multi-dimensional defect information of the industrial product, wherein the multi-dimensional defect information includes defect type information, defect location information, and defect shape information. The output control module is used to perform different levels of early warning processing based on the multidimensional defect information and generate corresponding quality inspection reports.
2. The system according to claim 1, characterized in that, The detection and reasoning module is used to perform multi-dimensional defect detection based on the industrial product image using a trained first defect detection model, and obtain the defect type information of the industrial product. Based on the image of the industrial product, multidimensional defect detection is performed using a trained second defect detection model to obtain the defect location information of the industrial product. Based on the image of the industrial product, multidimensional defect detection is performed using a trained third defect detection model to obtain the defect shape information of the industrial product.
3. The system according to claim 2, characterized in that, The first defect detection model is a ResNet network model, the second defect detection model is a YOLO network model, and the third defect detection model is a Mask R-CNN network model.
4. The system according to claim 2, characterized in that, The system includes a neural network training module; The neural network training module is used to train the first defect detection model, the second defect detection model, and the third defect detection model respectively based on the constructed training dataset, so as to obtain the trained defect detection model.
5. The system according to claim 4, characterized in that, The system includes a data annotation module; The data annotation module is used to perform multi-type label annotation on the collected industrial product images to construct a training dataset. The multi-type label annotation includes defect type annotation, defect location annotation, and defect shape annotation.
6. The system according to claim 1, characterized in that, The output control module is used to determine the level of a first warning message for the industrial product on the production line based on the defect type information; to determine the level of a second warning message for the industrial product on the production line based on the defect location information; and to determine the level of a third warning message for the industrial product on the production line based on the defect shape information; wherein the warning message levels include critical, severe, moderate, and minor levels. By combining the first warning information, the second warning information, and the level of the second warning information, different levels of warning processing are triggered, and corresponding quality inspection reports are generated.
7. The system according to claim 6, characterized in that, The output control module is used to integrate the first warning information, the second warning information, and the level of the second warning information to trigger different levels of warning processing: If any of the first warning message, the second warning message, or the third warning message contains a critical warning message, the production line operation shall be stopped immediately and emergency procedures shall be initiated. If any of the first warning information, the second warning information, or the third warning information is a severe warning information, the operation of the production line shall be suspended for rectification, and it shall be put back into operation after the rectification is completed. If only a general or minor level warning message exists among the first warning message, the second warning message, and the third warning message, then there is no need to stop the operation of the production line and optimize the automatic tracking and recording.
8. The system according to claim 1, characterized in that, The system includes a model deployment module; The model deployment module is used to compress and optimize the trained defect detection model through model quantization and model pruning, so as to deploy it in edge devices on the production line.
9. The system according to claim 1, characterized in that, The system includes a data preprocessing module; The data preprocessing module is used to preprocess the industrial product images acquired in real time by the image acquisition module on the production line to obtain preprocessed industrial product images. The preprocessing includes noise reduction, brightness normalization, image enhancement, and size cropping.
10. An industrial visual defect detection method, characterized in that, The method is performed based on the system according to any one of claims 1 to 9, and the method includes: Real-time image acquisition of industrial products on the production line; Based on the image of the industrial product, multidimensional defect detection is performed using a trained defect detection model to obtain multidimensional defect information of the industrial product, wherein the multidimensional defect information includes defect type information, defect location information, and defect shape information. Based on the multidimensional defect information, different levels of early warning processing are performed, and corresponding quality inspection reports are generated.