Electronic product quality inspection and analysis system based on image analysis

Abstract

The present invention relates to the technical field of product quality detection, and specifically discloses an electronic product quality detection and analysis system based on image analysis. The first acquisition module of the present invention collects product appearance and internal image feature information to lay the foundation for detection. Subsequently, the second acquisition module extracts surface cracks and oxidation area features based on the appearance image, and calculates the appearance feature index; the third acquisition module mines the internal infrared thermal imaging and stratification characteristic information; the fourth acquisition module generates a temperature feature index based on the infrared thermal imaging information; and the fifth acquisition module calculates the internal feature index based on the internal stratification information. Furthermore, the sixth acquisition module integrates the appearance, temperature, and internal feature indices to obtain a comprehensive product quality index. Finally, the analysis module conducts a comprehensive quantitative analysis of the product quality based on the index, determines the quality level, provides decision-making support for quality control, and can solve the problem of subdividing and quantitatively analyzing the quality level of electronic products.

Description

Technical Field

[0001] The present invention relates to the technical field of product quality detection, and in particular to an electronic product quality detection and analysis system based on image analysis. Background Art

[0002] Electronic product quality testing refers to the process of comprehensively evaluating the quality of electronic products through systematic technical means and standard processes. Its core goal is to ensure that the products meet the requirements of product design after they are produced. However, for ordinary consumers, the quality of PCB boards in electronic products is not easy to judge.

[0003] When ordinary consumers have PCB board quality rights protection, they usually need to send the disputed products for inspection. During the existing electronic product (PCB board) quality inspection process, it is usually necessary to use a CCD camera to take pictures of the electronic products for comparison and to identify quality problems of the electronic products. However, the existing photo comparison method can only determine whether there is a problem, but cannot perform a detailed quantitative analysis of the quality level of the electronic products. Therefore, an electronic product quality inspection and analysis system based on image analysis is needed to solve the above problems. Summary of the Invention

[0004] The purpose of the present invention is to provide an electronic product quality detection and analysis system based on image analysis to solve the technical problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] An electronic product quality detection and analysis system based on image analysis includes a first acquisition module, a second acquisition module, a third acquisition module, a fourth acquisition module, a fifth acquisition module, a sixth acquisition module and an analysis module:

[0007] A first acquisition module is configured to acquire image feature information of the electronic product, wherein the image feature information includes appearance image feature information and internal image feature information;

[0008] a second acquisition module, configured to acquire surface crack feature information and oxidized region feature information according to the appearance image feature information, and acquire an appearance feature index according to the surface crack feature information and the oxidized region feature information;

[0009] a third acquisition module, configured to acquire internal infrared thermal imaging feature information and internal layering characteristic information based on the internal image feature information;

[0010] a fourth acquisition module, configured to acquire infrared thermal imaging sequence information according to the internal infrared thermal imaging characteristic information, and acquire a temperature characteristic index according to the infrared thermal imaging sequence information;

[0011] a fifth acquisition module, configured to acquire layer length information and layer gap information according to the internal layer characteristic information, and acquire an internal characteristic index according to the layer length information and the layer gap information;

[0012] a sixth acquisition module, configured to acquire a comprehensive product quality index based on the appearance characteristic index, the temperature characteristic index, and the internal characteristic index;

[0013] The analysis module is used to perform comprehensive quantitative analysis on the quality of electronic products according to the comprehensive product quality index.

[0014] Preferably, the second acquisition module is further configured to perform grayscale processing on the appearance image feature information to obtain an appearance grayscale image;

[0015] The second acquisition module is further used to segment the appearance grayscale image according to grayscale color depth based on Canny edge detection to obtain multiple dark grayscale image regions, and use pixel feature information of the dark grayscale image regions as surface crack feature information;

[0016] The second acquisition module is further configured to sequentially acquire the number of pixels of the corresponding dark grayscale image regions according to the plurality of dark grayscale image regions, and accumulate the plurality of pixel numbers to obtain a first total pixel number;

[0017] The second acquisition module is further configured to assign the first total number of pixels according to a preset pixel number-area list to obtain the area of ​​the crack region;

[0018] The second acquisition module is further configured to acquire a plurality of pixel pairs of the appearance grayscale image based on the gray level co-occurrence matrix, and acquire corresponding brightness differences of a plurality of pixel pairs according to the plurality of pixel pairs;

[0019] The second acquisition module is further used to sequentially determine whether the brightness difference of the plurality of pixel pairs is greater than a preset threshold;

[0020] If the brightness difference of the pixel pair is greater than a preset threshold, the brightness differences of the pixel pairs greater than the preset threshold are counted to obtain a plurality of pixel pair brightness differences, and a corresponding plurality of first pixels are obtained based on the plurality of pixel pair brightness differences, and a corresponding second total number of pixels is obtained based on the plurality of first pixels;

[0021] The second acquisition module is further configured to assign a value to the second total number of pixels according to a preset pixel number-area list to obtain an area of ​​the oxidized region, and use the regional characteristic information corresponding to the area of ​​the oxidized region as the oxidized region characteristic information;

[0022] The second acquisition module is further used to calculate an appearance influence coefficient according to the area of ​​the crack region, the area of ​​the oxidized region, and the surface area of ​​a preset electronic product, and use the appearance influence coefficient as an appearance feature index.

[0023] Preferably, the fourth acquisition module is further configured to acquire infrared radiation energy distribution information within a preset time period based on the internal infrared thermal imaging characteristic information, and acquire a thermal imaging grayscale image based on the infrared radiation energy distribution information;

[0024] a fourth acquisition module, further configured to acquire a heat diffusion path according to the thermal imaging grayscale image, and acquire a plurality of imaging grayscale pixels according to the heat diffusion path;

[0025] A fourth acquisition module is further configured to assign radiation intensity values ​​to the plurality of imaging grayscale pixels based on a preset imaging grayscale pixel-radiation intensity value to obtain a plurality of radiation intensity values;

[0026] The fourth acquisition module is further configured to sort the plurality of radiation intensity values ​​based on a preset time sequence to obtain radiation intensity sequence information, and use the radiation intensity sequence information as infrared thermal imaging sequence information;

[0027] a fourth acquisition module, further configured to acquire a radiation intensity value variation curve within the electronic product according to the infrared thermal imaging sequence information, and acquire a plurality of radiation intensity peak values ​​according to the radiation intensity value variation curve;

[0028] The fourth acquisition module is further used to obtain an average radiation intensity peak value based on the multiple radiation intensity peak values, and convert the average radiation intensity peak value into a temperature value based on the Stefan-Boltzmann law, and use the temperature value as a temperature characteristic index.

[0029] Preferably, the fifth acquisition module is further configured to acquire a side internal image of the electronic product based on the internal image feature information, and perform grayscale processing on the side internal image to obtain a side internal grayscale image, wherein the side internal grayscale image includes a first light grayscale image area and a first dark grayscale image area;

[0030] A fifth acquisition module is further configured to acquire a plurality of segmentation edge curves of the first dark grayscale image region based on Canny edge detection, and sort the plurality of segmentation edge curves from top to bottom to obtain a segmentation edge curve sorting table;

[0031] a fifth acquisition module, further configured to filter the plurality of segmentation edge curves according to a maximum length value to obtain a first segmentation edge curve, map the first segmentation edge curve to a two-dimensional coordinate system, obtain first coordinates and second coordinates at both ends of the first segmentation edge curve, calculate a coordinate distance between the first coordinate and the second coordinate based on Euclidean distance, and use length information corresponding to the coordinate distance as layered length information;

[0032] A fifth acquisition module is further configured to acquire a layer gap distance between every two segmentation edge curves according to the segmentation edge curve sorting table;

[0033] a fifth acquiring module, further configured to filter the plurality of layer gap intervals according to a maximum interval value to obtain a first layer gap interval, and use gap information corresponding to the first layer gap interval as layer gap information;

[0034] The fifth acquisition module is further configured to perform weighted calculation on the first layer gap distance and the first layer gap distance to obtain an internal comprehensive damage coefficient, and use the internal comprehensive damage coefficient as an internal characteristic index.

[0035] Preferably, the sixth acquisition module is further configured to acquire an appearance feature index vector according to the appearance feature index, and perform normalization processing on the appearance feature index vector to obtain a normalized value of the appearance feature index vector;

[0036] a sixth acquisition module, further configured to acquire a temperature characteristic index vector according to the temperature characteristic index, and normalize the temperature characteristic index vector to obtain a normalized value of the temperature characteristic index vector;

[0037] a sixth acquisition module, further configured to acquire an internal characteristic index vector according to the internal characteristic index, and perform normalization processing on the internal characteristic index vector to obtain a normalized value of the internal characteristic index vector;

[0038] The sixth acquisition module is further used to perform weighted quantitative calculation on the normalized value of the appearance characteristic index vector, the normalized value of the temperature characteristic index vector and the normalized value of the internal characteristic index vector to obtain a comprehensive product quality index.

[0039] Preferably, the analysis module is further configured to obtain a plurality of historical electronic product detection databases;

[0040] The analysis module is further configured to perform similarity matching between the current electronic product and the plurality of historical electronic products based on a cosine similarity model to obtain matching historical electronic products, obtain corresponding historical product comprehensive quality indexes based on the matching historical electronic products, and obtain a mean of the product comprehensive quality index based on the historical product comprehensive quality indexes and the product comprehensive quality index;

[0041] The analysis module is further used to determine whether the average value of the comprehensive quality index of the product is within a preset threshold range;

[0042] If the average value of the product comprehensive quality index is not within the preset threshold range and is greater than the maximum value of the preset threshold range, the electronic product quality corresponding to the average value of the product comprehensive quality index is determined to be excellent;

[0043] If the average value of the product comprehensive quality index is within the preset threshold range, the electronic product quality corresponding to the average value of the product comprehensive quality index is determined to be qualified;

[0044] If the average value of the product comprehensive quality index is not within the preset threshold range and is less than the minimum value of the preset threshold range, the quality of the electronic product corresponding to the average value of the product comprehensive quality index is determined to be unqualified.

[0045] This application also provides an electronic product quality detection and analysis method based on image analysis, including:

[0046] Acquiring image feature information of the electronic product, wherein the image feature information includes appearance image feature information and internal image feature information;

[0047] Acquire surface crack feature information and oxidized region feature information according to the appearance image feature information, and acquire an appearance feature index according to the surface crack feature information and the oxidized region feature information;

[0048] Acquire internal infrared thermal imaging feature information and internal layering characteristic information according to the internal image feature information;

[0049] Acquiring infrared thermal imaging sequence information according to the internal infrared thermal imaging characteristic information, and acquiring a temperature characteristic index according to the infrared thermal imaging sequence information;

[0050] Acquire layer length information and layer gap information according to the internal layer characteristic information, and acquire an internal characteristic index according to the layer length information and the layer gap information;

[0051] Obtaining a comprehensive product quality index based on the appearance characteristic index, temperature characteristic index, and internal characteristic index;

[0052] Conduct a comprehensive quantitative analysis of electronic product quality based on the product comprehensive quality index.

[0053] As a preference, it includes:

[0054] The step of obtaining surface crack feature information and oxidation region feature information according to the appearance image feature information, and obtaining an appearance feature index according to the surface crack feature information and the oxidation region feature information, comprises:

[0055] Performing grayscale processing on the appearance image feature information to obtain an appearance grayscale image;

[0056] Based on Canny edge detection, the appearance grayscale image is segmented according to the grayscale color depth to obtain multiple dark grayscale image regions, and the pixel feature information of the dark grayscale image regions is used as the surface crack feature information;

[0057] sequentially acquiring pixel numbers of corresponding dark grayscale image regions according to the plurality of dark grayscale image regions, and accumulating the plurality of pixel numbers to obtain a first total pixel number;

[0058] Assigning the first total number of pixels according to a preset pixel number-area list to obtain the area of ​​the crack region;

[0059] Acquire a plurality of pixel pairs of the appearance grayscale image based on the gray level co-occurrence matrix, and acquire corresponding brightness differences of a plurality of pixel pairs according to the plurality of pixel pairs;

[0060] sequentially determining whether a brightness difference between the plurality of pixel pairs is greater than a preset threshold;

[0061] If the brightness difference of the pixel pair is greater than a preset threshold, the brightness differences of the pixel pairs greater than the preset threshold are counted to obtain a plurality of pixel pair brightness differences, and a corresponding plurality of first pixels are obtained based on the plurality of pixel pair brightness differences, and a corresponding second total number of pixels is obtained based on the plurality of first pixels;

[0062] Assigning the second total number of pixels according to a preset pixel number-area list to obtain an oxidized region area, and using region feature information corresponding to the oxidized region area as oxidized region feature information;

[0063] Calculating an appearance influence coefficient based on the area of ​​the crack region, the area of ​​the oxidized region, and the surface area of ​​a preset electronic product;

[0064] The appearance influence coefficient is used as the appearance feature index.

[0065] The present application also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above system when executing the computer program.

[0066] The present application also provides a computer-readable storage medium having a computer program stored thereon, which implements the steps of the above system when executed by a processor.

[0067] The beneficial effects of the present application are as follows: the present invention constructs a systematic evaluation scheme around the quality inspection of electronic products. First, the first acquisition module collects product appearance and internal image feature information to lay the foundation for inspection. Subsequently, the second acquisition module extracts surface cracks and oxidation area features based on the appearance image, and calculates the appearance feature index; the third acquisition module mines the internal infrared thermal imaging and stratification characteristic information; the fourth acquisition module generates a temperature feature index based on the infrared thermal imaging information; the fifth acquisition module calculates the internal feature index based on the internal stratification information. Furthermore, the sixth acquisition module integrates the appearance, temperature, and internal feature indices to derive a comprehensive product quality index. Finally, the analysis module conducts a comprehensive quantitative analysis of product quality based on the index, determines the quality level, provides decision-making support for quality control, and can solve the problem of subdividing and quantitatively analyzing the quality level of electronic products. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 This is a schematic diagram of a system flow chart of an embodiment of the present application.

[0069] Figure 2 This is a schematic diagram of the method structure of an embodiment of the present application.

[0070] Figure 3 This is a schematic diagram of the internal structure of a computer device according to an embodiment of the present application.

[0071] The realization of the objectives, functional features and advantages of this application will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0072] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0073] like Figure 1-3 As shown, the present application provides an electronic product quality detection and analysis system based on image analysis, including a first acquisition module, a second acquisition module, a third acquisition module, a fourth acquisition module, a fifth acquisition module, a sixth acquisition module and an analysis module:

[0074] A first acquisition module is configured to acquire image feature information of the electronic product, wherein the image feature information includes appearance image feature information and internal image feature information;

[0075] a second acquisition module, configured to acquire surface crack feature information and oxidized region feature information according to the appearance image feature information, and acquire an appearance feature index according to the surface crack feature information and the oxidized region feature information;

[0076] a third acquisition module, configured to acquire internal infrared thermal imaging feature information and internal layering characteristic information based on the internal image feature information;

[0077] a fourth acquisition module, configured to acquire infrared thermal imaging sequence information according to the internal infrared thermal imaging characteristic information, and acquire a temperature characteristic index according to the infrared thermal imaging sequence information;

[0078] a fifth acquisition module, configured to acquire layer length information and layer gap information according to the internal layer characteristic information, and acquire an internal characteristic index according to the layer length information and the layer gap information;

[0079] a sixth acquisition module, configured to acquire a comprehensive product quality index based on the appearance characteristic index, the temperature characteristic index, and the internal characteristic index;

[0080] The analysis module is used to perform comprehensive quantitative analysis on the quality of electronic products according to the comprehensive product quality index.

[0081] As mentioned above, when ordinary consumers assert their rights regarding PCB board quality, they must submit the disputed product for inspection. Existing quality inspections of electronic products (PCB boards) typically require using a CCD camera to photograph and compare the electronic product, thereby identifying any quality issues. However, existing photographic comparison methods can only determine whether a problem exists, but are unable to provide a detailed, quantitative analysis of the product's quality. Therefore, the present invention first acquires image feature information of the electronic product through a first acquisition module. This image feature information includes both external and internal image feature information. This comprehensive collection of both the external and internal image features of the electronic product provides raw data support for subsequent inspection and analysis, forming the foundation of the entire inspection process. This image feature information can reflect the actual condition of the product. The external image can reveal whether there are surface defects, while the internal image can help identify potential internal issues. Furthermore, to implement image analysis-based quality inspection, relevant image data must first be acquired. Different types of image features contain a wealth of product quality information. Only by obtaining this information can the product quality be evaluated from various angles. Then, the second acquisition module obtains surface crack feature information and oxidation area feature information based on the appearance image feature information, and obtains the appearance feature index based on the surface crack feature information and the oxidation area feature information, and accurately identifies the cracks and oxidation areas in the appearance image, and quantifies the degree of influence of the appearance on the product quality by calculating the appearance feature index. The cracks and oxidation of the appearance may affect the appearance, protective performance and service life of the product. This step can provide an important basis for judging product quality. At the same time, appearance defects are one of the important factors affecting the quality of electronic products. Acquiring and quantifying these features through specific image processing algorithms can make the appearance quality assessment more scientific and accurate, thereby comprehensively evaluating the product quality. Then, the third acquisition module obtains the internal infrared thermal imaging feature information and internal stratification feature information based on the internal image feature information. This in-depth exploration of the thermal imaging and stratification characteristics inside the electronic product will help to discover possible thermal anomalies and stratification defects inside.Internal thermal anomalies may affect the performance stability of the product, and delamination defects may lead to a decrease in structural strength. This information is crucial for evaluating the internal quality of the product. Obtaining internal infrared thermal imaging and delamination characteristic information can evaluate product quality from an internal level and ensure product reliability. Next, the fourth acquisition module is used to obtain infrared thermal imaging sequence information based on the internal infrared thermal imaging characteristic information, and obtain a temperature characteristic index based on the infrared thermal imaging sequence information. In this way, the temperature characteristic index can quantify the thermal conditions inside the product. Among them, temperature is an important indicator reflecting the performance and stability of electronic products. Abnormal temperature may indicate that the product has potential faults. The temperature characteristic index helps to more intuitively evaluate the thermal performance of the product. At the same time, the infrared thermal imaging sequence information can dynamically reflect the thermal changes inside the product. Converting it into a temperature characteristic index can more conveniently quantify the internal thermal performance of the product, providing strong support for quality assessment. Next, the fifth acquisition module is used to obtain delamination length information and delamination gap information based on the internal delamination characteristic information, and obtain an internal characteristic index based on the delamination length information and the delamination gap information. The internal delamination length and gap are accurately measured, and the internal characteristic index is calculated to quantify the impact of internal structural defects on product quality. Delamination can affect the mechanical and electrical properties of electronic products. This step helps assess the integrity and reliability of the product's internal structure. Internal delamination is a key internal defect in electronic products (PCBs). Obtaining delamination length and gap information and quantifying it into an internal characteristic index can more accurately assess the impact of internal structural defects on product quality, providing a key basis for quality assessment. Secondly, the sixth acquisition module is used to obtain a comprehensive product quality index based on the appearance characteristic index, temperature characteristic index, and internal characteristic index. This comprehensively considers multiple factors, including appearance, thermal performance, and internal structure, to produce a comprehensive index that fully reflects product quality. The comprehensive product quality index makes quality assessment more comprehensive and objective, avoiding the limitations of single-metric assessment. Finally, the analysis module is used to perform a comprehensive quantitative analysis of electronic product quality based on the comprehensive product quality index. This allows for accurate determination of the quality level of electronic products, providing a basis for quantitative product quality analysis, offering decision support for classification management, and addressing the issue of segmented quantitative analysis of electronic product quality.

[0082] In one embodiment, the second acquisition module is further configured to perform grayscale processing on the appearance image feature information to obtain an appearance grayscale image;

[0083] The second acquisition module is further used to segment the appearance grayscale image according to grayscale color depth based on Canny edge detection to obtain multiple dark grayscale image regions, and use pixel feature information of the dark grayscale image regions as surface crack feature information;

[0084] The second acquisition module is further configured to sequentially acquire the number of pixels of the corresponding dark grayscale image regions according to the plurality of dark grayscale image regions, and accumulate the plurality of pixel numbers to obtain a first total pixel number;

[0085] The second acquisition module is further configured to assign the first total number of pixels according to a preset pixel number-area list to obtain the area of ​​the crack region;

[0086] The second acquisition module is further configured to acquire a plurality of pixel pairs of the appearance grayscale image based on the gray level co-occurrence matrix, and acquire corresponding brightness differences of a plurality of pixel pairs according to the plurality of pixel pairs;

[0087] The second acquisition module is further used to sequentially determine whether the brightness difference of the plurality of pixel pairs is greater than a preset threshold;

[0088] If the brightness difference of the pixel pair is greater than a preset threshold, the brightness differences of the pixel pairs greater than the preset threshold are counted to obtain a plurality of pixel pair brightness differences, and a corresponding plurality of first pixels are obtained based on the plurality of pixel pair brightness differences, and a corresponding second total number of pixels is obtained based on the plurality of first pixels;

[0089] The second acquisition module is further configured to assign a value to the second total number of pixels according to a preset pixel number-area list to obtain an area of ​​the oxidized region, and use the regional characteristic information corresponding to the area of ​​the oxidized region as the oxidized region characteristic information;

[0090] The second acquisition module is further configured to calculate an appearance influence coefficient based on the crack area, the oxidation area, and the surface area of ​​the preset electronic product, wherein the calculation formula is:

[0091] ;

[0092] in, represents the appearance influence coefficient, represents the crack area, represents the area of ​​oxidation region; Indicates the surface area of ​​the preset electronic product. The first weight representing the ratio of the crack area to the surface area of ​​the preset electronic product, A second weight representing the ratio of the area of ​​the oxidized region to the surface area of ​​the preset electronic product.

[0093] As described above, the second acquisition module of the present invention is also used to grayscale the appearance image feature information to obtain an appearance grayscale image. This converts the color appearance image into a grayscale image, simplifies the image information, and highlights the image's brightness features, facilitating subsequent edge detection and feature extraction operations. Because in a grayscale image, pixel values ​​only represent brightness, this reduces interference from color information, making it easier for subsequent processing to detect key features in the image, such as cracks. The appearance grayscale image is then segmented based on grayscale color depth based on Canny edge detection to obtain multiple dark grayscale image regions. The pixel feature information of the dark grayscale image regions is then used as surface crack feature information. In this way, the Canny edge detection algorithm is used to accurately identify areas in the appearance grayscale image where cracks may exist. These dark grayscale image regions are considered potential crack regions, and their pixel feature information is obtained as surface crack features, providing a basis for determining the presence and characteristics of cracks. This helps intuitively determine the location and approximate shape of the crack. Secondly, the pixel counts of the corresponding dark grayscale image regions are sequentially obtained based on the multiple dark grayscale image regions, and the multiple pixel counts are accumulated to obtain a first total pixel count. By counting the pixel counts of the potential crack region, the size of the crack region is quantified. The first total pixel count can reflect the scale of the crack region. This is important for assessing the impact of cracks on product appearance quality. A larger number of pixels generally means a larger crack region, which has a greater impact on product quality. Furthermore, in quantitative analysis, specific data is required to describe the size of the crack region. The pixel count is an intuitive and easy-to-calculate quantitative indicator. By accumulating the pixel counts of the dark grayscale image regions, the scale of the crack region can be accurately obtained, providing data support for the subsequent calculation of the crack region area and appearance characteristic index. The first total pixel count is then assigned according to a preset pixel count-area table to obtain the crack region area. This conversion of pixel counts into actual area makes the size of the crack region more intuitive and practical. A preset pixel count-area table establishes a correspondence between pixel count and actual area. This conversion allows for a more accurate assessment of the impact of cracked areas on product appearance, facilitating comprehensive analysis with other indicators. Multiple pixel pairs of the appearance grayscale image are then obtained based on the grayscale co-occurrence matrix. Based on these pixel pairs, the corresponding brightness differences between the multiple pixel pairs are then obtained. This grayscale co-occurrence matrix can reflect the spatial relationship and grayscale distribution between pixels in the image. By obtaining the brightness differences between these pixel pairs, the degree of variation between different pixels in the image can be analyzed, allowing identification of areas with significant grayscale variation within the image, which may be associated with oxidized areas.This helps locate potential oxidation areas. Oxidation areas often manifest as abnormal grayscale distribution in images. The grayscale co-occurrence matrix is ​​an effective texture analysis tool. By calculating the brightness differences between pixel pairs, it can mine texture features in the image and identify characteristic information related to oxidation areas, providing a basis for subsequent identification of oxidation areas. Next, the brightness differences between multiple pixel pairs are determined to be greater than a preset threshold. If the brightness differences between pixel pairs are greater than the preset threshold, the brightness differences between pixel pairs that are greater than the preset threshold are counted to obtain multiple pixel pair brightness differences. Based on these multiple pixel pair brightness differences, multiple first pixels corresponding to the differences are obtained, and then a second total number of pixels corresponding to the multiple first pixels is obtained. By setting a preset threshold (because surface roughness associated with surface oxidation is greater than that of a normal surface, which is reflected by darker pixels), pixel pairs with significant grayscale variations are screened out. These pixel pairs are likely to be oxidation areas. The number of pixels corresponding to these pixel pairs, i.e., the second total number of pixels, is counted to quantify the size of the oxidation area. This helps accurately determine the size of the oxidation zone, providing data for assessing the impact of oxidation on the product's appearance quality. Finally, the appearance impact coefficient is calculated based on the area of ​​the cracked area, the area of ​​the oxidized area, and the surface area of ​​the pre-determined electronic product. This comprehensively considers the area ratios of the cracked and oxidized areas to calculate the appearance impact coefficient, comprehensively quantifying the impact of appearance defects on product quality. The appearance impact coefficient integrates multiple appearance defect factors into a single quantitative indicator, making it convenient to evaluate the overall quality of a product in conjunction with other quality indicators (such as the temperature characteristic index and the internal characteristic index). The impact of cracks and oxidation areas on product appearance quality cannot be viewed in isolation; their proportions of the total product surface area must be considered comprehensively. The appearance influence coefficient is calculated using this formula, where the weight coefficient generally ranges from 0 to 1. For example, if the impact of cracks on product quality is considered to be more serious than oxidation, then the weight coefficient of cracks is set higher. In this solution, the corresponding weights are set by the crack and oxidation areas, which can more scientifically evaluate the comprehensive impact of appearance defects on product quality, providing an important basis for the final product quality assessment. At the same time, the appearance influence coefficient is used as the appearance characteristic index, which can provide a clear quantitative representation of appearance quality in the comprehensive product quality assessment, facilitating subsequent calculations and quality grade determination.

[0094] In one embodiment, the fourth acquisition module is further configured to acquire infrared radiation energy distribution information within a preset time period based on the internal infrared thermal imaging characteristic information, and acquire a thermal imaging grayscale image based on the infrared radiation energy distribution information;

[0095] a fourth acquisition module, further configured to acquire a heat diffusion path according to the thermal imaging grayscale image, and acquire a plurality of imaging grayscale pixels according to the heat diffusion path;

[0096] A fourth acquisition module is further configured to assign radiation intensity values ​​to the plurality of imaging grayscale pixels based on a preset imaging grayscale pixel-radiation intensity value to obtain a plurality of radiation intensity values;

[0097] The fourth acquisition module is further configured to sort the plurality of radiation intensity values ​​based on a preset time sequence to obtain radiation intensity sequence information, and use the radiation intensity sequence information as infrared thermal imaging sequence information;

[0098] a fourth acquisition module, further configured to acquire a radiation intensity value variation curve within the electronic product according to the infrared thermal imaging sequence information, and acquire a plurality of radiation intensity peak values ​​according to the radiation intensity value variation curve;

[0099] The fourth acquisition module is further used to obtain an average radiation intensity peak value based on the multiple radiation intensity peak values, and convert the average radiation intensity peak value into a temperature value based on the Stefan-Boltzmann law, and use the temperature value as a temperature characteristic index.

[0100] As described above, the fourth acquisition module of the present invention is further configured to obtain infrared radiation energy distribution information within a preset time period based on the internal infrared thermal imaging characteristic information, and to obtain a thermal imaging grayscale image based on the infrared radiation energy distribution information. This acquisition of infrared radiation energy distribution information and its conversion into a thermal imaging grayscale image presents the heat source information within the electronic product as a visual grayscale image. The thermal imaging grayscale image can intuitively reflect temperature differences between different regions within the product, with different grayscale values ​​corresponding to different temperatures. This facilitates subsequent analysis of heat diffusion paths and temperature changes, providing a basis for identifying potential thermal anomalies. Furthermore, the thermal imaging grayscale image is used to obtain a heat diffusion path, and multiple imaging grayscale pixels are obtained based on the heat diffusion path. This analysis of the thermal imaging grayscale image determines the heat diffusion path, which can reveal the direction and distribution of heat transfer within the product. The multiple imaging grayscale pixels obtained contain information about the heat source along the heat diffusion path, providing specific data points for subsequent acquisition of radiation intensity values. The heat diffusion path is a key element in studying the internal thermal performance of electronic products. Understanding the heat diffusion path can reveal areas of heat concentration or abnormal heat transfer paths. The purpose of acquiring imaging grayscale pixels is to more accurately quantify the heat source information along the heat diffusion path, providing a specific data foundation for subsequent analysis and making the study of thermal characteristics more targeted and accurate. Subsequently, based on the preset imaging grayscale pixel-radiation intensity value, multiple imaging grayscale pixels are assigned radiation intensity values ​​to obtain multiple radiation intensity values. This allows each imaging grayscale pixel to be assigned a radiation intensity value, converting the pixel information in the thermal imaging grayscale image into radiation intensity data. The radiation intensity value can more accurately quantify the degree of thermal radiation in the area represented by each pixel, facilitating subsequent precise analysis of the thermal radiation situation and providing a quantitative basis for evaluating the internal thermal performance of the product. At the same time, the imaging grayscale pixel itself is only a visual representation of thermal information. By assigning a radiation intensity value, it is converted into physically meaningful quantitative data. In this way, when analyzing thermal performance, more accurate calculations and comparisons can be made based on radiation intensity values, improving the accuracy and scientific nature of thermal performance evaluation. Multiple radiation intensity values ​​are then sorted based on a preset time series to obtain radiation intensity sequence information, which is then used as infrared thermal imaging sequence information (obtained during power-on testing of the PCB). Sorting the radiation intensity values ​​according to the preset time series forms radiation intensity sequence information, i.e., infrared thermal imaging sequence information. This sequence information can reflect how thermal radiation within the product changes over time, displaying the dynamic process of thermal radiation. This helps discover changing trends and abnormal fluctuations in thermal radiation, and provides a basis for analyzing the thermal performance of the product at different time points. A curve of radiation intensity values ​​within the electronic product is obtained based on the infrared thermal imaging sequence information, and multiple radiation intensity peaks are obtained based on the curve of radiation intensity values. This converts the infrared thermal imaging sequence information into a curve of radiation intensity values, graphically and more intuitively displaying the changing trend of thermal radiation intensity over time.By analyzing the curve, multiple radiation intensity peaks are obtained. These peaks represent the maximum internal thermal radiation of the product. These peaks may be associated with key heat-generating components or abnormal thermal phenomena within the product, helping to identify potential thermal issues and critical heat-generating areas. The curve also provides a more intuitive view of the data's changing trends, facilitating observation and analysis. Radiation intensity peaks are key characteristic points in thermal radiation variations. Extracting these peaks allows for more sensitive detection of thermal anomalies within the product, providing important clues for quality inspection and analysis, helping to determine whether there are potential issues such as overheating within the product. Finally, the average radiation intensity peak is calculated from these multiple peaks. This average is converted to a temperature value based on the Stefan-Boltzmann law, which serves as a temperature characteristic index. This calculation comprehensively considers the characteristics of multiple peaks, yielding a single value that represents the overall level of internal thermal radiation within the product. This value is converted to a temperature value based on the Stefan-Boltzmann law, which states that the total energy radiated per unit area of ​​a blackbody per unit time is proportional to the fourth power of the blackbody's thermodynamic temperature. This allows thermal radiation information to be converted into a straightforward temperature indicator. The temperature characteristic index quantifies the thermal conditions within a product, making it convenient to evaluate the overall product quality alongside other quality indicators (such as the appearance characteristic index and the internal characteristic index). It can be used to determine whether the product is operating within the normal temperature range. Meanwhile, the peak radiation intensity reflects the maximum value of thermal radiation, but there are differences between multiple peaks. Calculating the average radiation intensity peak provides a more representative value. Converting this into a temperature value and using it as the temperature characteristic index facilitates comprehensive analysis with other indicators within a unified quality assessment system, enabling accurate assessment of product quality.

[0101] In one embodiment, the fifth acquisition module is further configured to acquire a side internal image of the electronic product based on the internal image feature information, and perform grayscale processing on the side internal image to obtain a side internal grayscale image, wherein the side internal grayscale image includes a first light grayscale image region and a first dark grayscale image region;

[0102] A fifth acquisition module is further configured to acquire a plurality of segmentation edge curves of the first dark grayscale image region based on Canny edge detection, and sort the plurality of segmentation edge curves from top to bottom to obtain a segmentation edge curve sorting table;

[0103] a fifth acquisition module, further configured to filter the plurality of segmentation edge curves according to a maximum length value to obtain a first segmentation edge curve, map the first segmentation edge curve to a two-dimensional coordinate system, obtain first coordinates and second coordinates at both ends of the first segmentation edge curve, calculate a coordinate distance between the first coordinate and the second coordinate based on Euclidean distance, and use length information corresponding to the coordinate distance as layered length information;

[0104] A fifth acquisition module is further configured to acquire a layer gap distance between every two segmentation edge curves according to the segmentation edge curve sorting table;

[0105] a fifth acquiring module, further configured to filter the plurality of layer gap intervals according to a maximum interval value to obtain a first layer gap interval, and use gap information corresponding to the first layer gap interval as layer gap information;

[0106] The fifth acquisition module is further configured to perform weighted calculation on the first layer gap distance and the first layer gap distance to obtain an internal comprehensive damage coefficient, and use the internal comprehensive damage coefficient as an internal characteristic index.

[0107] As described above, the fifth acquisition module of the present invention is further configured to acquire a side internal image of the electronic product based on the internal image feature information, perform grayscale processing on the side internal image, and obtain a side internal grayscale image. The side internal grayscale image includes a first light grayscale image region and a first dark grayscale image region. Acquiring the side internal image and converting it into a grayscale image simplifies image information, highlights brightness differences in the internal structure, and facilitates subsequent identification and analysis of layered regions. Grayscale images can reduce color information interference, making the distinction between layered regions and other regions more distinct, and providing a clearer base image for detecting layered characteristics. Secondly, based on Canny edge detection, multiple segmentation edge curves are acquired for the first dark grayscale image region, and the multiple segmentation edge curves are sorted from top to bottom to obtain a segmentation edge curve sorting table. Thus, the Canny edge detection algorithm is used to accurately locate the edge curves of the first dark grayscale image region (typically corresponding to the layered region) in the side internal grayscale image, and the sorting process forms a segmentation edge curve sorting table. This helps clearly present the boundary information of the delamination area, facilitating subsequent measurement and analysis of delamination length and gaps, making the extraction of delamination features more systematic and accurate. Because longer lengths have a greater impact on quality, this solution uses the maximum length as a quantification criterion to provide a basis for unified quantification. Multiple segmentation edge curves are then filtered according to the maximum length value to obtain a first segmentation edge curve. This first segmentation edge curve is mapped to a two-dimensional coordinate system, and the first and second coordinates of the two ends of the first segmentation edge curve are obtained. The coordinate distance between the first and second coordinates is calculated based on the Euclidean distance, and the length information corresponding to the coordinate distance is used as the delamination length information. In this way, the segmentation edge curve with the maximum length is selected to represent the main delamination situation. By mapping to a two-dimensional coordinate system and calculating the coordinate distance, the delamination length can be accurately measured. Delamination length is an important indicator for assessing the severity of internal delamination. Longer delaminations may have a greater impact on product structure and performance, providing a key basis for judging the quality of the product's internal structure. Among multiple segmentation edge curves, the curve with the maximum length often represents the most representative delamination situation and can reflect the main delamination defects. Using a two-dimensional coordinate system and Euclidean distance to calculate length, edge information in the image can be converted into specific length values, achieving precise quantification of delamination lengths, facilitating subsequent quality assessment and comparison. The delamination gap spacing between each pair of segmentation edge curves is then determined according to the segmentation edge curve sorting table. This provides a comprehensive understanding of the spacing between delamination areas. Delamination gap spacing is an important parameter for measuring delamination characteristics. Different gap spacings can affect the electrical and mechanical properties of the product, facilitating a more detailed assessment of the impact of internal delamination on product quality. In addition to delamination length, delamination gap spacing also has an impact on product quality.By measuring the distance between each pair of segmented edge curves, the distribution characteristics of the delamination area can be obtained, providing richer data for comprehensive assessment of internal delamination conditions and making quality assessment more comprehensive and accurate. Secondly, multiple delamination gaps are filtered according to the maximum gap value to obtain the first delamination gap distance. The gap information corresponding to the first delamination gap distance is used as the delamination gap information. This selects the largest delamination gap distance as the representative value, highlighting the most serious delamination gap situation. The maximum delamination gap distance can be a key factor affecting product performance. Using it as delamination gap information can more intuitively reflect the potential risk of internal delamination to product quality, facilitating focused attention and assessment. Furthermore, among multiple delamination gap distances, the maximum gap distance often has the most significant impact on product quality. By selecting the maximum gap distance and using it as delamination gap information, key issues can be identified, the analysis process can be simplified, and potential quality risk points can be more effectively identified in quality assessment. Finally, the delamination length information and the first delamination gap distance are weighted to obtain the internal comprehensive damage coefficient, which is used as the internal characteristic index. This comprehensive internal damage coefficient is calculated by comprehensively considering the two key factors of delamination length and delamination gap distance. This index comprehensively quantifies the impact of internal delamination on product quality, integrating multiple delamination-related factors into a single value. This facilitates the assessment of a product's overall quality alongside other quality indicators (such as the appearance characteristic index and the temperature characteristic index). Both delamination length and delamination gap spacing significantly impact the quality of a product's internal structure, and considering any one factor alone cannot fully assess the impact of delamination. Using a weighted calculation approach, we can comprehensively consider the importance of different factors to form a unified internal characteristic index, providing a more scientific and comprehensive basis for product quality assessment.

[0108] In one embodiment, the sixth acquisition module is further configured to acquire an appearance feature index vector according to the appearance feature index, and perform normalization processing on the appearance feature index vector to obtain a normalized value of the appearance feature index vector;

[0109] a sixth acquisition module, further configured to acquire a temperature characteristic index vector according to the temperature characteristic index, and normalize the temperature characteristic index vector to obtain a normalized value of the temperature characteristic index vector;

[0110] a sixth acquisition module, further configured to acquire an internal characteristic index vector according to the internal characteristic index, and perform normalization processing on the internal characteristic index vector to obtain a normalized value of the internal characteristic index vector;

[0111] The sixth acquisition module is further used to perform weighted quantitative calculation on the normalized value of the appearance characteristic index vector, the normalized value of the temperature characteristic index vector and the normalized value of the internal characteristic index vector to obtain a comprehensive product quality index.

[0112] As described above, the sixth acquisition module of the present invention is further configured to obtain an appearance feature index vector based on the appearance feature index and normalize the appearance feature index vector to obtain a normalized value of the appearance feature index vector. This converts the appearance feature index into vector form, making the appearance quality information more suitable for mathematical operations and model processing. Normalization maps the value of the appearance feature index vector to a specific interval (typically [0, 1]), eliminating the impact of differences in dimensionality and numerical range between different feature indices. This ensures that in subsequent comprehensive calculations, the various feature indices can be fairly compared and reasonably integrated under a unified standard, thereby improving the accuracy and comparability of the calculation results. A temperature feature index vector is then obtained based on the temperature feature index and normalized to obtain a normalized value of the temperature feature index vector. Similar to the first step, converting the temperature feature index into vector form facilitates subsequent mathematical operations. The normalized normalized value of the temperature feature index vector eliminates the differences in dimensionality and numerical range between the temperature feature index and other feature indices. In this way, when comprehensively evaluating product quality, the temperature characteristic index can be considered on an equal footing with other characteristic indices, accurately reflecting the impact of a product's thermal performance on its overall quality and avoiding assessment bias due to numerical differences. Secondly, an internal characteristic index vector is obtained based on the internal characteristic index, and the internal characteristic index vector is normalized to obtain a normalized internal characteristic index vector value. Converting the internal characteristic index into a vector and normalizing it is also intended to ensure comparability and fairness in comprehensive calculations. The internal characteristic index reflects the quality of the electronic product's internal structure. Through vector representation and normalization, internal structure information can be integrated into a unified calculation framework, accurately measuring the impact of the internal structure on the overall product quality and avoiding unreasonable amplification or reduction of the internal characteristic index in comprehensive evaluations due to numerical differences. Finally, the normalized appearance characteristic index vector value, the normalized temperature characteristic index vector value, and the normalized internal characteristic index vector value are weighted and quantified to obtain a comprehensive product quality index. This weighted quantification comprehensively considers the three factors of appearance, temperature, and internal structure, integrating quality information from different aspects into a single value: the comprehensive product quality index. Weighted calculation can assign weights to different factors based on their importance to product quality, making the calculation results more accurately reflect the actual quality level of the product, providing an intuitive and comprehensive quantitative indicator for the comprehensive evaluation of product quality. The quality of electronic products is determined by multiple aspects. Appearance, thermal performance, and internal structure are all important factors affecting product quality, but the degree of their impact on product quality may vary. Using weighted quantitative calculation, reasonable weights can be assigned to the normalized values ​​of different characteristic index vectors based on actual conditions, thereby more scientifically integrating various factors to derive a comprehensive quality index that accurately reflects the overall quality of the product, meeting the needs of comprehensive product quality assessment in actual production.

[0113] In one embodiment, the analysis module is further configured to obtain a plurality of historical electronic product detection databases;

[0114] The analysis module is further configured to perform similarity matching between the current electronic product and the plurality of historical electronic products based on a cosine similarity model to obtain matching historical electronic products, obtain corresponding historical product comprehensive quality indexes based on the matching historical electronic products, and obtain a mean of the product comprehensive quality index based on the historical product comprehensive quality indexes and the product comprehensive quality index;

[0115] The analysis module is further used to determine whether the average value of the comprehensive quality index of the product is within a preset threshold range;

[0116] If the average value of the product comprehensive quality index is not within the preset threshold range and is greater than the maximum value of the preset threshold range, the electronic product quality corresponding to the average value of the product comprehensive quality index is determined to be excellent;

[0117] If the average value of the product comprehensive quality index is within the preset threshold range, the electronic product quality corresponding to the average value of the product comprehensive quality index is determined to be qualified;

[0118] If the average value of the product comprehensive quality index is not within the preset threshold range and is less than the minimum value of the preset threshold range, the quality of the electronic product corresponding to the average value of the product comprehensive quality index is determined to be unqualified.

[0119] As described above, the analysis module of the present invention is used to obtain a plurality of historical electronic product detection databases, thereby collecting a large amount of historical electronic product detection data to form a detection database. The database contains quality-related information of products from different periods and different batches, providing a rich reference sample for subsequent comparative analysis. Through these historical data, the fluctuation range and change trend of product quality can be understood, providing a historical basis for judging the quality of current products, and enhancing the reliability and accuracy of quality assessment. At the same time, when evaluating the quality of current electronic products, it is difficult to accurately judge their quality level based solely on the various characteristic indexes of a single product. The historical detection database provides multiple sets of data that reflect the quality of products under different production conditions and environments. Comparing the current product with historical data can evaluate product quality from a more macro perspective, discover potential problems and development trends in product quality, and then perform similarity matching between the current electronic product and multiple historical electronic products based on the cosine similarity model to obtain matching historical electronic products. Based on the matching historical electronic products, the corresponding historical product comprehensive quality index is obtained, and the average of the product comprehensive quality index is obtained based on the historical product comprehensive quality index and the product comprehensive quality index. In this way, using the cosine similarity model, historical products similar to the current electronic product are found in the historical detection database to obtain their historical product comprehensive quality indexes. By calculating the average comprehensive quality index of the current product and the matching historical products, the quality of the current product and similar historical products can be comprehensively considered, the uncertainty of a single product data can be reduced, and the quality level of the current product can be evaluated more comprehensively and objectively. Finally, it is determined whether the average comprehensive quality index of the product is within the preset threshold range. If the average comprehensive quality index of the product is not within the preset threshold range and is greater than the preset threshold range, the quality of the electronic product corresponding to the average comprehensive quality index of the product is determined to be excellent. If the average comprehensive quality index of the product is within the preset threshold range, the quality of the electronic product corresponding to the average comprehensive quality index of the product is determined to be qualified. If the average comprehensive quality index of the product is not within the preset threshold range and is less than the minimum value of the preset threshold range, the quality of the electronic product corresponding to the average comprehensive quality index of the product is determined to be unqualified. In this way, the average comprehensive quality index of the product is judged according to the preset threshold range, and the quality of the electronic product is divided into three levels: excellent, qualified, and unqualified. This quantitative quality judgment method makes the quality assessment results more intuitive and clear, making it easier for enterprises to classify and manage products and take corresponding measures for products of different quality levels, such as recognizing and promoting excellent products and improving or eliminating unqualified products. At the same time, in actual production, it is necessary to clearly divide product quality into different levels for management and decision-making.The preset threshold range is set based on the company's quality standards and production requirements. By comparing the mean value of the product's comprehensive quality index with the threshold range, the product quality level can be quickly and accurately judged, the efficiency and operability of quality inspection can be improved, the company's needs for product quality control can be met, and the problem of subdividing and quantitatively analyzing the quality level of electronic products can be solved.

[0120] This application also provides an electronic product quality detection and analysis method based on image analysis, including:

[0121] Acquiring image feature information of the electronic product, wherein the image feature information includes appearance image feature information and internal image feature information;

[0122] Acquire surface crack feature information and oxidized region feature information according to the appearance image feature information, and acquire an appearance feature index according to the surface crack feature information and the oxidized region feature information;

[0123] Acquire internal infrared thermal imaging feature information and internal layering characteristic information according to the internal image feature information;

[0124] Acquiring infrared thermal imaging sequence information according to the internal infrared thermal imaging characteristic information, and acquiring a temperature characteristic index according to the infrared thermal imaging sequence information;

[0125] Acquire layer length information and layer gap information according to the internal layer characteristic information, and acquire an internal characteristic index according to the layer length information and the layer gap information;

[0126] Obtaining a comprehensive product quality index based on the appearance characteristic index, temperature characteristic index, and internal characteristic index;

[0127] Conduct a comprehensive quantitative analysis of electronic product quality based on the product comprehensive quality index.

[0128] In one embodiment, the method includes:

[0129] The step of obtaining surface crack feature information and oxidation region feature information according to the appearance image feature information, and obtaining an appearance feature index according to the surface crack feature information and the oxidation region feature information, comprises:

[0130] Performing grayscale processing on the appearance image feature information to obtain an appearance grayscale image;

[0131] Based on Canny edge detection, the appearance grayscale image is segmented according to the grayscale color depth to obtain multiple dark grayscale image regions, and the pixel feature information of the dark grayscale image regions is used as the surface crack feature information;

[0132] sequentially acquiring pixel numbers of corresponding dark grayscale image regions according to the plurality of dark grayscale image regions, and accumulating the plurality of pixel numbers to obtain a first total pixel number;

[0133] Assigning the first total number of pixels according to a preset pixel number-area list to obtain the area of ​​the crack region;

[0134] Acquire a plurality of pixel pairs of the appearance grayscale image based on the gray level co-occurrence matrix, and acquire corresponding brightness differences of a plurality of pixel pairs according to the plurality of pixel pairs;

[0135] sequentially determining whether a brightness difference between the plurality of pixel pairs is greater than a preset threshold;

[0136] If the brightness difference of the pixel pair is greater than a preset threshold, the brightness differences of the pixel pairs greater than the preset threshold are counted to obtain a plurality of pixel pair brightness differences, and a corresponding plurality of first pixels are obtained based on the plurality of pixel pair brightness differences, and a corresponding second total number of pixels is obtained based on the plurality of first pixels;

[0137] Assigning the second total number of pixels according to a preset pixel number-area list to obtain an oxidized region area, and using region feature information corresponding to the oxidized region area as oxidized region feature information;

[0138] Calculating an appearance influence coefficient based on the area of ​​the crack region, the area of ​​the oxidized region, and the surface area of ​​a preset electronic product;

[0139] The appearance influence coefficient is used as the appearance feature index.

[0140] The present application also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above system when executing the computer program.

[0141] The present application also provides a computer-readable storage medium having a computer program stored thereon, which implements the steps of the above system when executed by a processor.

[0142] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media provided in this application and used in the embodiments may 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 many forms such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct RAM bus dynamic RAM (DRDRAM), and RAM bus dynamic RAM (RDRAM).

[0143] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, apparatus, article, or method comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, apparatus, article, or method. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, apparatus, article, or method comprising the element.

[0144] The above description is only a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. An electronic product quality inspection and analysis system based on image analysis, characterized in that: It includes a first acquisition module, which is used to acquire appearance image feature information and internal image feature information of the electronic product; a second acquisition module, configured to acquire surface crack feature information and oxidized region feature information according to the appearance image feature information, and acquire an appearance feature index according to the surface crack feature information and the oxidized region feature information; a third acquisition module, configured to acquire internal infrared thermal imaging feature information and internal layering characteristic information based on the internal image feature information; a fourth acquisition module, configured to acquire infrared thermal imaging sequence information according to the internal infrared thermal imaging characteristic information, and acquire a temperature characteristic index according to the infrared thermal imaging sequence information; a fifth acquisition module, configured to acquire delamination length information and delamination gap information based on the internal delamination characteristic information, and to acquire an internal characteristic index based on the delamination length information and the delamination gap information, wherein the module is further configured to acquire a side internal image of the electronic product based on the internal image characteristic information, perform grayscale processing on the side internal image to obtain a side internal grayscale image, wherein the side internal grayscale image includes a first light grayscale image region and a first dark grayscale image region, acquire multiple segmentation edge curves in the first dark grayscale image region, sort the multiple segmentation edge curves from top to bottom to obtain a segmentation edge curve sorting table, filter the multiple segmentation edge curves to obtain a first segmentation edge curve, acquire first and second coordinates at both ends of the first segmentation edge curve, and a coordinate distance between the first and second coordinates as delamination length information, acquire a delamination gap spacing between every two delamination edge curves according to the delamination edge curve sorting table, filter the multiple delamination gap spacings to obtain a first delamination gap spacing as delamination gap information, perform weighted calculation on the delamination length information and the first delamination gap spacing to obtain an internal comprehensive damage coefficient as the internal characteristic index; a sixth acquisition module, configured to acquire a comprehensive product quality index based on the appearance characteristic index, the temperature characteristic index, and the internal characteristic index; The analysis module is used to perform comprehensive quantitative analysis on the quality of electronic products according to the comprehensive product quality index.

2. The electronic product quality detection and analysis system based on image analysis according to claim 1, characterized in that: The second acquisition module is also used to perform grayscale processing on the appearance image feature information to obtain an appearance grayscale image, and to segment the appearance grayscale image to obtain multiple dark grayscale image areas as surface crack feature information, obtain a first total pixel number based on the multiple dark grayscale image areas, assign a value to the first total pixel number table to obtain the crack area, obtain the brightness difference of multiple pixel pairs of the appearance grayscale image, and judge in turn whether the brightness difference of the multiple pixel pairs is greater than a preset threshold. If it is greater than the preset threshold, the brightness difference of the pixel pairs greater than the preset threshold is counted to obtain multiple pixel pair brightness differences, and obtain corresponding multiple first pixels based on the multiple pixel pair brightness differences, and obtain corresponding second total pixel numbers based on the multiple first pixels, assign a value to the second total pixel number to obtain the oxidation area, and use it as the oxidation area feature information, calculate the appearance influence coefficient based on the crack area area, the oxidation area, and the surface area of ​​the preset electronic product, and use it as the appearance feature index.

3. The electronic product quality detection and analysis system based on image analysis according to claim 1, characterized in that: The fourth acquisition module is also used to obtain a thermal imaging grayscale image based on the internal infrared thermal imaging characteristic information, obtain multiple imaging grayscale pixel points based on the thermal imaging grayscale image, assign radiation intensity to the multiple imaging grayscale pixel points to obtain multiple radiation intensity values, sort the multiple radiation intensity values ​​to obtain radiation intensity sequence information, and use it as infrared thermal imaging sequence information, obtain a radiation intensity value change curve in the electronic product based on the infrared thermal imaging sequence information, and obtain multiple radiation intensity peaks based on the radiation intensity value change curve, obtain an average radiation intensity peak value based on the multiple radiation intensity peaks, and convert the average radiation intensity peak value into a temperature value based on the Stefan-Boltzmann law, and use the temperature value as a temperature characteristic index.

4. The electronic product quality inspection and analysis system based on image analysis according to claim 1, characterized in that: The sixth acquisition module is also used to obtain an appearance feature index vector based on the appearance feature index, and normalize the appearance feature index vector to obtain a normalized value of the appearance feature index vector, obtain a temperature feature index vector based on the temperature feature index, and normalize the temperature feature index vector to obtain a normalized value of the temperature feature index vector, obtain an internal feature index vector based on the internal feature index, and normalize the internal feature index vector to obtain a normalized value of the internal feature index vector, and perform weighted quantitative calculation on the normalized value of the appearance feature index vector, the normalized value of the temperature feature index vector and the normalized value of the internal feature index vector to obtain a comprehensive product quality index.

5. The electronic product quality inspection and analysis system based on image analysis according to claim 1, characterized in that: The analysis module is also used to obtain a detection database of multiple historical electronic products, perform similarity matching between the current electronic product and the multiple historical electronic products to obtain matching historical electronic products, and obtain the corresponding historical product comprehensive quality index based on the matching historical electronic products, and obtain the product comprehensive quality index mean based on the historical product comprehensive quality index and the product comprehensive quality index, and judge whether the product comprehensive quality index mean is within a preset threshold range. If the product comprehensive quality index mean is not within the preset threshold range and is greater than the maximum value of the preset threshold range, then the quality of the electronic product corresponding to the product comprehensive quality index mean is judged to be excellent. If the product comprehensive quality index mean is within the preset threshold range, then the quality of the electronic product corresponding to the product comprehensive quality index mean is judged to be qualified. If the product comprehensive quality index mean is not within the preset threshold range and is less than the minimum value of the preset threshold range, then the quality of the electronic product corresponding to the product comprehensive quality index mean is judged to be unqualified.

6. An electronic product quality detection and analysis method based on image analysis, used to execute the electronic product quality detection and analysis system based on image analysis according to any one of claims 1 to 5, characterized in that: include: Acquiring image feature information of the electronic product, wherein the image feature information includes appearance image feature information and internal image feature information; Acquire surface crack feature information and oxidized region feature information according to the appearance image feature information, and acquire an appearance feature index according to the surface crack feature information and the oxidized region feature information; Acquire internal infrared thermal imaging feature information and internal layering characteristic information according to the internal image feature information; Acquiring infrared thermal imaging sequence information according to the internal infrared thermal imaging characteristic information, and acquiring a temperature characteristic index according to the infrared thermal imaging sequence information; Acquire layer length information and layer gap information according to the internal layer characteristic information, and acquire an internal characteristic index according to the layer length information and the layer gap information; Obtaining a comprehensive product quality index based on the appearance characteristic index, temperature characteristic index, and internal characteristic index; Conduct a comprehensive quantitative analysis of electronic product quality based on the product comprehensive quality index.

7. The electronic product quality detection and analysis method based on image analysis according to claim 6, characterized in that: include: The step of obtaining surface crack feature information and oxidation region feature information according to the appearance image feature information, and obtaining an appearance feature index according to the surface crack feature information and the oxidation region feature information, comprises: Performing grayscale processing on the appearance image feature information to obtain an appearance grayscale image; Based on Canny edge detection, the appearance grayscale image is segmented according to the grayscale color depth to obtain multiple dark grayscale image regions, and the pixel feature information of the dark grayscale image regions is used as the surface crack feature information; sequentially acquiring pixel numbers of corresponding dark grayscale image regions according to the plurality of dark grayscale image regions, and accumulating the plurality of pixel numbers to obtain a first total pixel number; Assigning the first total number of pixels according to a preset pixel number-area list to obtain the area of ​​the crack region; Acquire a plurality of pixel pairs of the appearance grayscale image based on the gray level co-occurrence matrix, and acquire corresponding brightness differences of a plurality of pixel pairs according to the plurality of pixel pairs; sequentially determining whether a brightness difference between the plurality of pixel pairs is greater than a preset threshold; If the brightness difference of the pixel pair is greater than a preset threshold, the brightness differences of the pixel pairs greater than the preset threshold are counted to obtain a plurality of pixel pair brightness differences, and a corresponding plurality of first pixels are obtained based on the plurality of pixel pair brightness differences, and a corresponding second total number of pixels is obtained based on the plurality of first pixels; Assigning the second total number of pixels according to a preset pixel number-area list to obtain an oxidized region area, and using region feature information corresponding to the oxidized region area as oxidized region feature information; Calculating an appearance influence coefficient based on the area of ​​the crack region, the area of ​​the oxidized region, and the surface area of ​​a preset electronic product; The appearance influence coefficient is used as the appearance feature index.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the system according to any one of claims 1 to 5 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the system according to any one of claims 1 to 5 are implemented.

Images