Industrial product surface defect rapid detection and classification method
Through automated detection methods that integrate high-resolution images, spectral and depth data, the problem of inefficiency of traditional detection methods is solved, efficient, accurate detection and intelligent traceability of surface defects of industrial products are achieved, and the quality control capabilities of large-scale production are improved.
Patent Information
- Application Number
- CN202510278773.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-10
- Publication Date
- 2025-07-11
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Traditional industrial products surface defect detection methods are inefficient, prone to misjudgment and missed inspection, difficult to meet the needs of efficient and accurate inspection of large-scale production, and lack intelligent re-inspection and defect cause traceability mechanisms.
The fusion of high-resolution image data, spectral data and depth data is adopted, combined with image enhancement, optical correction and three-dimensional reconstruction, and the product analysis data is automatically generated, and the output detection tag is compared with the set threshold value, and the rechecking threshold is set to realize automated rechecking and defect cause traceability.
It realizes efficient and accurate detection of surface defects of industrial products, reduces false inspections and missed inspections, improves detection efficiency and accuracy, and provides stable quality control support.
Smart Images
Figure CN120293983A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of product defect detection, and particularly to a method for rapid detection and classification of surface defects of industrial products. Background Art
[0002] In modern industrial production, product quality control is an important link to ensure market competitiveness and customer satisfaction. However, with the diversification of product types and the complexity of manufacturing processes, the detection of surface defects of industrial products faces increasing challenges. Traditional detection methods mainly rely on manual inspection or single-sensor devices, which are not only inefficient but also prone to misjudgment and missed detection. In particular, some subtle texture defects, minor color differences, or irregular surface unevenness are difficult to accurately identify and quantify in traditional image detection methods. Therefore, how to achieve efficient, accurate, and comprehensive surface defect detection has become an urgent problem to be solved.
[0003] With the continuous improvement of the automation level of industrial production lines, the requirements for detection speed in large-scale production are becoming increasingly stringent. Relying solely on manual or single-device detection not only fails to adapt to the high-speed operation of the production line but also leads to a lag in the feedback of quality problems, thereby increasing the defective rate and affecting production efficiency. The differences in materials and processes among different products make it difficult for a single detection method to take into account multi-dimensional quality requirements. Therefore, a detection method that integrates multiple data sources is needed to achieve rapid and accurate detection and classification of product surface defects to meet the quality management needs in a large-scale production environment.
[0004] In addition, the current ability to re-inspect detected abnormal products is limited. When occasional or marginal abnormalities occur, they are often misjudged as defective products, resulting in waste of resources. The lack of a perfect re-inspection mechanism not only reduces the stability of the detection system but also causes unnecessary rework and increased costs. In addition, the tracing of defect causes mostly relies on manual analysis, lacking intelligent data retrieval and comparison functions, and unable to find the root cause of problems in a timely and accurate manner. This makes the discovery and handling of product quality problems lack systematicness and timeliness, further exacerbating the management difficulty in the production process. Summary of the Invention
[0005] To solve the above technical problems, the present invention provides a method for rapid detection and classification of surface defects of industrial products.
[0006] A method for rapid detection and classification of surface defects of industrial products, the method comprising: Collecting surface data of industrial products and performing preprocessing to generate product analysis data; Compare the product analysis data with the established product analysis standards to output product analysis labels, where the product analysis labels include product detection defect labels, product detection anomaly labels, and product detection normal labels; Re-detect the products with product detection anomaly labels generated to obtain the product analysis labels after re-inspection; Retrieve the data of the products with product detection defect labels generated to obtain the product analysis data with defects, and trace the defect causes based on the product analysis data with defects.
[0007] Further, the industrial product surface data includes high-resolution image data, spectral data, and depth data; among them, the high-resolution image data includes product surface texture and obvious defects, the spectral data includes product materials and product color differences, and the depth data includes the number of product surface concave points and the number of product surface convex points; Let the product color difference be β, and calculate the product color difference β according to the collected product brightness L, product red-green component α, and product blue-green component γ. The logical formula for the product color difference β is: β = ; where, is the preset product brightness comparison standard value, is the preset product red-green component comparison standard value, is the product blue-green component comparison standard value.
[0008] Further, the logic for collecting and preprocessing industrial product surface data is: Perform image enhancement and noise filtering on the high-resolution image data, and extract the processed product surface texture and obvious defect information; perform optical correction and noise reduction on the spectral data, remove the ambient light interference and extract the processed product materials and product color differences; perform denoising and three-dimensional reconstruction on the depth data, and re-count to obtain the processed number of product surface concave points and the number of product surface convex points; output the processed product surface texture, obvious defect information, product materials, product color differences, the number of product surface concave points, and the number of product surface convex points as product analysis data.
[0009] Further, the product analysis standards include texture feature thresholds, defect size thresholds, spectral deviation thresholds, color difference tolerance thresholds, concave point number thresholds, and convex point number thresholds.
[0010] Further, the logic for comparing the product analysis data with the established product analysis standards to output product analysis labels is: Compare the surface texture, obvious defect information, product material, product color difference, the number of concave points on the product surface, and the number of convex points on the product surface with the established texture feature thresholds, defect size thresholds, spectral deviation thresholds, color difference tolerance thresholds, concave point number thresholds, and convex point number thresholds. If more than one product analysis data exceeds the standard, generate a product detection defect label; if one product analysis data exceeds the standard, generate a product detection anomaly label; if no product analysis data exceeds the standard, generate a product detection normal label.
[0011] Further, the logic for re-detecting the generated product detection anomaly label to obtain the re-inspection result is as follows: Perform data re-inspection on the product analysis data of the product with the generated product detection anomaly label, and set the re-inspection anomaly count threshold. If the number of repeated re-inspections is greater than or equal to the re-inspection anomaly count threshold, generate a product detection defect label for this product. If the number of repeated re-inspections is less than the re-inspection anomaly count threshold and the product analysis label of this product is detected as a product detection normal label, then generate a product detection normal label for this product; if the number of repeated re-inspections is less than the re-inspection anomaly count threshold and the product analysis label of this product is detected as a product detection defect label, then generate a product detection defect label for this product.
[0012] Further, the logic for retrieving data of the generated product detection defect label to obtain the product analysis data with defects and tracing the cause of the defect based on the product analysis data with defects is as follows: After the product is marked as "product detection defect label", retrieve various defect problems related to the product analysis data with defects from the database, and compare with the parameters of the product analysis data of other products without defects. Select the defect problem with the most suitable parameters and output it as the cause of the defect.
[0013] An industrial product surface defect rapid detection and classification system for executing any industrial product surface defect rapid detection and classification method, the system includes: Data acquisition module: used to collect industrial product surface data and perform preprocessing to generate product analysis data; Product analysis module: used to compare the product analysis data with the established product analysis standards to output product analysis labels, and the product analysis labels include product detection defect labels, product detection anomaly labels, and product detection normal labels; Defect re-inspection module: used to re-detect the products with the generated product detection anomaly labels to obtain the re-inspected product analysis labels; Defect Traceability Module: It is used to retrieve data of the generated product detection defect labels to obtain product analysis data of defective products, and trace the defect causes based on the product analysis data of defective products.
[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: Through the fusion of high-resolution image data, spectral data, and depth data, the present invention realizes the efficient detection of the surface of industrial products. The high-resolution image data captures the surface texture and obvious defects of the product from multiple angles. The spectral data makes up for the subtle defects that cannot be recognized by traditional images. The depth data accurately detects the spatial defects of the uneven surface. After preprocessing such as automated image enhancement, light correction, and three-dimensional reconstruction of the data, product analysis data including product surface texture, obvious defects, product materials, color difference, number of concave points, and number of convex points is quickly generated. And through comparison based on the set texture feature threshold, defect size threshold, spectral deviation threshold, color difference tolerance threshold, number threshold of concave points and convex points, product detection defect labels, detection anomaly labels, or detection normal labels are automatically output to ensure the efficiency and accuracy of the detection process. In addition, for the generated detection anomaly labels, a retest anomaly count threshold is set, and the retest results dynamically update the product detection labels according to the count threshold to avoid misdetection or missed detection caused by accidental factors. At the same time, data of the generated detection defect labels is retrieved, and by comparing the analysis data of different products, the most likely defect causes are traced. This solution effectively improves the efficiency and accuracy of industrial product detection through full-process automation and intelligent defect traceability, and provides efficient and stable quality control support for large-scale production environments. Description of the Drawings
[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required to be used in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.
[0016] Figure 1 It is a flowchart of a method for rapid detection and classification of surface defects of industrial products provided in Embodiment 1 of the present invention; Figure 2 It is a module diagram of a system for rapid detection and classification of surface defects of industrial products provided in Embodiment 2 of the present invention. To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.
[0017] Embodiment 1 Please refer to Figure 1 As shown, this embodiment discloses and provides a method for rapid detection and classification of surface defects of industrial products. The method includes: S110: Collect surface data of industrial products and perform preprocessing to generate product analysis data; Specifically, the surface data of the industrial products includes high-resolution image data, spectral data, and depth data. Among them, the high-resolution image data includes product surface texture and obvious defects, the spectral data includes product materials and product color differences, and the depth data includes the number of concave points on the product surface and the number of convex points on the product surface; Let the product color difference be β. According to the collected product brightness L, product red-green component α, and product blue-green component γ, the product color difference β is calculated. The logical formula for the product color difference β is: β = ; Among them, is a preset product brightness comparison standard value, is a preset product red-green component comparison standard value, is the product blue-green component comparison standard value; It should be noted that L represents the brightness (or lightness) component of the color, with a value range of 0 to 100. The larger the L value, the brighter the color; α represents the red and green components of the color, with a value range of -128 to +127, and α represents the yellow and blue components of the color, with a value range of -128 to +127; It should be noted that: The product surface texture and obvious defects are captured from multiple angles by a high-resolution camera. The spectral data collects the spectral reflection information of the product surface through a spectral sensor to make up for the subtle defects that cannot be recognized by traditional image data; the depth data collects the three-dimensional topography information of the product surface through a laser or structured light device, and the depth data can accurately detect the spatial defects of the uneven surface of the product; Specifically, the logic for collecting surface data of industrial products and performing preprocessing is: Perform image enhancement and noise filtering on high-resolution image data to extract the processed product surface texture and obvious defect information; perform light correction and noise reduction on spectral data to remove ambient light interference and extract the processed product material and product color difference; perform denoising and three-dimensional reconstruction on depth data to recalculate and obtain the number of concave points and the number of convex points on the processed product surface; output the processed product surface texture, obvious defect information, product material, product color difference, the number of concave points on the product surface, and the number of convex points on the product surface as product analysis data.
[0018] S120: Compare the product analysis data with the established product analysis standards to output product analysis labels, where the product analysis labels include product detection defect labels, product detection anomaly labels, and product detection normal labels; Specifically, the product analysis standards include texture feature thresholds, defect size thresholds, spectral deviation thresholds, color difference tolerance thresholds, concave point number thresholds, and convex point number thresholds; It should be noted that: the texture feature threshold is used to analyze the normal range of the surface texture; the defect size threshold is used to evaluate the size of obvious defects; the spectral deviation threshold is used to compare the spectral characteristics of the product material; the color difference tolerance threshold is used to determine whether the color difference exceeds the standard; the concave point number threshold and the convex point number threshold are respectively used to limit the number of unevenness on the product surface.
[0019] Specifically, the logic of comparing the product analysis data with the established product analysis standards to output product analysis labels is as follows: Compare the product surface texture, obvious defect information, product material, product color difference, the number of concave points on the product surface, and the number of convex points on the product surface with the established texture feature thresholds, defect size thresholds, spectral deviation thresholds, color difference tolerance thresholds, concave point number thresholds, and convex point number thresholds. If more than one item of product analysis data exceeds the standard, generate a product detection defect label; if one item of product analysis data exceeds the standard, generate a product detection anomaly label; if no product analysis data exceeds the standard, generate a product detection normal label.
[0020] S130: Retest the products with product detection anomaly labels to obtain the retested product analysis labels; Specifically, the logic of retesting the products with product detection anomaly labels to obtain the retest results is as follows: Perform data recheck on the product analysis data for generating product detection abnormal labels, set the threshold of the number of recheck abnormalities. If the number of repeated rechecks is greater than or equal to the threshold of the number of recheck abnormalities, generate product detection defect labels for the product. If the number of repeated rechecks is less than the threshold of the number of recheck abnormalities and the product analysis label of the product is detected as a product detection normal label, then generate a product detection normal label for the product. If the number of repeated rechecks is less than the threshold of the number of recheck abnormalities and the product analysis label of the product is detected as a product detection defect label, then generate product detection defect labels for the product.
[0021] S140: Retrieve data for generating product detection defect labels to obtain product analysis data with defects, and trace the defect causes based on the product analysis data with defects. Specifically, the logic of retrieving data for generating product detection defect labels to obtain product analysis data with defects and tracing the defect causes based on the product analysis data with defects is as follows: After the product is marked as "product detection defect label", retrieve multiple types of defect problems of the product analysis data with defects from the database, and compare the parameters of the product analysis data of other products without defects based on the multiple types of defect problems. Select the defect problem with the most suitable parameters and output it as the defect cause.
[0022] Embodiment 2 Please refer to Figure 2 As shown in the figure, based on the unified inventive concept, this embodiment publicly provides an industrial product surface defect rapid detection and classification system, and the system includes: Data acquisition module S210: used to collect industrial product surface data and perform preprocessing to generate product analysis data. Specifically, the industrial product surface data includes high-resolution image data, spectral data, and depth data; among them, The high-resolution image data includes product surface texture and obvious defects, the spectral data includes product materials and product color differences, and the depth data includes the number of concave points on the product surface and the number of convex points on the product surface. Let the product color difference be β, and calculate the product color difference β according to the collected product brightness L, product red-green component α, and product blue-green component γ. The logical formula of the product color difference β is: β = ; Among them, is the preset product brightness comparison standard value, is the preset product red-green component comparison standard value, is the product blue-green component comparison standard value; Specifically, the logic for collecting and preprocessing the surface data of industrial products is as follows: Perform image enhancement and noise filtering on the high-resolution image data to extract the processed surface texture and obvious defect information of the product; perform light correction and noise reduction on the spectral data to remove ambient light interference and extract the processed product material and product color difference; denoise and perform three-dimensional reconstruction on the depth data to recalculate and obtain the number of concave points and the number of convex points on the processed product surface; output the processed product surface texture, obvious defect information, product material, product color difference, number of concave points on the product surface, and number of convex points on the product surface as product analysis data.
[0023] Product analysis module S220: used to compare the product analysis data with established product analysis standards to output product analysis labels, where the product analysis labels include product detection defect labels, product detection anomaly labels, and product detection normal labels; Specifically, the product analysis standards include texture feature thresholds, defect size thresholds, spectral deviation thresholds, color difference tolerance thresholds, concave point number thresholds, and convex point number thresholds; Specifically, the logic for comparing the product analysis data with established product analysis standards to output product analysis labels is as follows: Compare the product surface texture, obvious defect information, product material, product color difference, number of concave points on the product surface, and number of convex points on the product surface with the established texture feature thresholds, defect size thresholds, spectral deviation thresholds, color difference tolerance thresholds, concave point number thresholds, and convex point number thresholds. If more than one piece of product analysis data exceeds the standard, generate a product detection defect label; if one piece of product analysis data exceeds the standard, generate a product detection anomaly label; if no product analysis data exceeds the standard, generate a product detection normal label.
[0024] Defect re-inspection module S230: used to re-inspect the products with product detection anomaly labels to obtain the re-inspected product analysis labels; Specifically, the logic for re-inspecting the products with product detection anomaly labels to obtain the re-inspection results is as follows: Re-inspect the product analysis data of the products with product detection anomaly labels, set the re-inspection anomaly count threshold. If the number of repeated re-inspections is greater than or equal to the re-inspection anomaly count threshold, generate a product detection defect label for this product. If, when the number of repeated re-inspections is less than the re-inspection anomaly count threshold, it is detected that the product analysis label of this product is a product detection normal label, then generate a product detection normal label for this product; if, when the number of repeated re-inspections is less than the re-inspection anomaly count threshold, it is detected that the product analysis label of this product is a product detection defect label, then generate a product detection defect label for this product; Defect Traceability Module S240: Used to retrieve data of the generated product inspection defect labels to obtain product analysis data with defects, and trace the defect causes based on the product analysis data with defects; Specifically, the logic for retrieving data of the generated product inspection defect labels to obtain product analysis data with defects and tracing the defect causes based on the product analysis data with defects is as follows: After the product is marked with the "product inspection defect label", retrieve multiple types of defect problems of the product analysis data with defects from the database, and compare by retrieving the parameters of the product analysis data of other non-defective products based on the multiple types of defect problems. Select the defect problem with the most suitable parameters and output it as the defect cause.
[0025] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present invention are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website, computer, server, or data center to another website, computer, server, or data center via a wired network or a wireless network. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that includes one or more collections of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.
[0026] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in the present invention can be implemented by electronic hardware or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.
[0027] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.
[0028] In several embodiments provided by the present invention, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only for a method for rapid detection and classification of surface defects of industrial products. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces. The indirect couplings or communication connections of the devices or units can be in electrical, mechanical, or other forms.
[0029] The units described as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0030] In addition, in each embodiment of the present invention, the functional units can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit.
[0031] As described above, the above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention and should be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.
[0032] Finally: The above are only the preferred embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims
1. A method for rapid detection and classification of surface defects of industrial products, characterized in that, The method includes: Collecting surface data of industrial products and performing preprocessing to generate product analysis data; Comparing the product analysis data with established product analysis standards to output product analysis labels, where the product analysis labels include product detection defect labels, product detection anomaly labels, and product detection normal labels; Re-detecting the products with product detection anomaly labels to obtain the product analysis labels after re-inspection; Retrieving data of the products with product detection defect labels to obtain product analysis data with defects, and tracing the causes of the defects based on the product analysis data with defects.
2. The rapid detection and classification method for surface defects of industrial products according to claim 1, characterized in that, The surface data of the industrial products includes high-resolution image data, spectral data, and depth data; among them, the high-resolution image data includes product surface texture and obvious defects, the spectral data includes product materials and product color differences, and the depth data includes the number of concave points on the product surface and the number of convex points on the product surface; Let the product color difference be β, and the product color difference β is calculated according to the collected product brightness L, product red-green component α, and product blue-green component γ. The logical formula for the product color difference β is: β= ； Among them, is the preset standard value for product brightness comparison, is the preset standard value for product red-green component comparison, is the standard value for product blue-green component comparison.
3. A rapid detection and classification method for surface defects of industrial products according to claim 2, characterized in that, The logic for collecting surface data of industrial products and performing preprocessing is: Performing image enhancement and noise filtering on the high-resolution image data, and extracting the processed product surface texture and obvious defect information; Performing optical correction and noise reduction on the spectral data, removing ambient light interference and extracting the processed product materials and product color differences; through denoising and three-dimensional reconstruction of the depth data, re-counting to obtain the number of concave points on the processed product surface and the number of convex points on the product surface; Outputting the processed product surface texture, obvious defect information, product materials, product color differences, the number of concave points on the product surface, and the number of convex points on the product surface as product analysis data.
4. A method for rapid detection and classification of surface defects of industrial products according to claim 3, characterized in that, The product analysis standards include texture feature thresholds, defect size thresholds, spectral deviation thresholds, color difference tolerance thresholds, concave point number thresholds, and convex point number thresholds.
5. A rapid detection and classification method for surface defects of industrial products according to claim 4, characterized in that, The logic for comparing the product analysis data with established product analysis standards to output product analysis labels is: Comparing the product surface texture, obvious defect information, product materials, product color differences, the number of concave points on the product surface, and the number of convex points on the product surface with the established texture feature thresholds, defect size thresholds, spectral deviation thresholds, color difference tolerance thresholds, concave point number thresholds, and convex point number thresholds. If more than one item of product analysis data exceeds the standard, then generate product detection defect labels; if one item of product analysis data exceeds the standard, then generate product detection anomaly labels; if no product analysis data exceeds the standard, then generate product detection normal labels.
6. The rapid detection and classification method for surface defects of industrial products according to claim 5, characterized in that, The logic for re-detecting the products with product detection anomaly labels to obtain the re-inspection results is: Performing data re-inspection on the product analysis data of the products with product detection anomaly labels, setting a re-inspection anomaly count threshold. If the number of repeated re-inspections is greater than or equal to the re-inspection anomaly count threshold, then generate product detection defect labels for this product. If when the number of repeated re-inspections is less than the re-inspection anomaly count threshold and it is detected that the product analysis label of this product is a product detection normal label, then generate product detection normal labels for this product; If the product analysis label of the product is detected as a product detection defect label when the number of reciprocating re-inspections is less than the re-inspection anomaly count threshold, then a product detection defect label is generated for the product.
7. A method for rapid detection and classification of surface defects of industrial products according to claim 6, characterized in that, The logic for retrieving data for the generated product detection defect label to obtain product analysis data with defects and tracing the cause of the defect based on the product analysis data with defects is as follows: After the product is marked with the "product detection defect label", multiple types of defect problems related to the product analysis data with defects are retrieved from the database, and parameters of other product analysis data without defects are retrieved based on the multiple types of defect problems for comparison. The defect problem with the most suitable parameters is selected and output as the cause of the defect.
8. An industrial product surface defect rapid detection and classification system for implementing the industrial product surface defect rapid detection and classification method according to any one of claims 1-7, characterized in that, The system includes: A data acquisition module: used to collect surface data of industrial products and perform preprocessing to generate product analysis data; A product analysis module: used to compare the product analysis data with established product analysis standards to output a product analysis label, where the product analysis label includes a product detection defect label, a product detection anomaly label, and a product detection normal label; A defect re-inspection module: used to re-inspect products with a generated product detection anomaly label to obtain a product analysis label after re-inspection; A defect tracing module: used to retrieve data for the generated product detection defect label to obtain product analysis data with defects and trace the cause of the defect based on the product analysis data with defects.