Mobile power supply appearance quality inspection method based on machine vision

Abstract

The present invention relates to the technical field of appearance quality detection, and in particular to a method for detecting the appearance quality of a mobile power supply based on machine vision. The method comprises the following steps: using a high-resolution spectral camera to collect multi-angle spectral images of a mobile power supply, and performing texture structure reflection spectrum analysis to obtain a mobile power supply texture structure reflection spectrum; performing texture ductility analysis on the mobile power supply texture structure reflection spectrum, and performing geometric distortion analysis to obtain texture defect geometric distortion data; performing rough perception evaluation on the texture defect geometric distortion data to obtain texture defect rough perception data; performing defect severity level classification on the mobile power supply texture structure reflection spectrum to obtain a defect severity level classification report; performing automated quality monitoring firmware design based on the defect severity level classification report to obtain automated quality monitoring firmware. The present invention makes the appearance quality detection technology more perfect by optimizing the appearance quality detection technology.

Description

Technical Field

[0001] The present invention relates to the technical field of appearance quality detection, and in particular to a method for detecting the appearance quality of a mobile power source based on machine vision. Background Art

[0002] With the advancement of technology, mobile power banks, as portable charging devices, have been widely used to meet the charging needs of various electronic products. However, as the market pays more and more attention to product quality, the appearance quality inspection of mobile power banks has become increasingly important. Traditional manual inspection methods are not only time-consuming and labor-intensive, but also easily affected by human factors, resulting in inconsistent inspections or missing defects, making their inspection efficiency low and accuracy poor. Summary of the invention

[0003] Based on this, it is necessary to provide a mobile power supply appearance quality inspection method based on machine vision to solve at least one of the above technical problems.

[0004] To achieve the above purpose, a method for detecting the appearance quality of a mobile power supply based on machine vision is provided, the method comprising the following steps:

[0005] Step S1: using a high-resolution spectral camera to collect multi-angle spectral images of the mobile power supply to obtain a spectral sharpened image of the mobile power supply;

[0006] Step S2: performing texture structure reflection spectrum analysis on the mobile power source spectral sharpening image to obtain the mobile power source texture structure reflection spectrum; performing texture ductility analysis on the mobile power source texture structure reflection spectrum to obtain texture connection ductility data;

[0007] Step S3: performing geometric distortion analysis on the texture extension disconnection data to obtain texture defect geometric distortion data; performing roughness perception evaluation on the texture defect geometric distortion data to obtain texture defect roughness perception data;

[0008] Step S4: classifying the defect severity level of the mobile power supply texture structure reflection map according to the texture defect roughness perception data and the texture defect geometric distortion data to obtain a defect severity level classification report;

[0009] Step S5: Design an automated quality monitoring firmware based on the defect severity classification report to obtain the automated quality monitoring firmware, and send the automated quality monitoring firmware to the cloud platform to perform real-time appearance quality detection.

[0010] The present invention uses a high-resolution spectral camera to collect multi-angle spectral images of a mobile power supply, and can obtain spectrally sharpened images of the mobile power supply. These images reveal the reflection characteristics of the mobile power supply at different wavelengths, so that the subtle differences on the surface of the object can be captured more clearly. This spectral sharpening not only improves the clarity and contrast of the image, but also emphasizes subtle surface features, laying the foundation for subsequent texture analysis. By collecting data from multiple angles, a comprehensive understanding of the surface state of the mobile power supply can be ensured, avoiding potential defects that are missed due to a single viewing angle. The texture structure reflection spectrum analysis is performed on the acquired mobile power supply spectrally sharpened image, so as to obtain the texture structure reflection spectrum of the mobile power supply. This analysis process mainly focuses on identifying and parsing the texture pattern of the mobile power supply surface, including its reflection characteristics and the distribution of surface texture. Through the analysis of the texture structure reflection spectrum, the details of the surface texture and its ductility can be deeply understood. These data are further used for texture ductility analysis to help evaluate the behavior and coherence of the surface texture under different conditions, thereby providing a comprehensive understanding of the texture and laying a solid foundation for subsequent defect detection and quality control. The geometric distortion analysis of texture extension and disconnection data is performed to obtain the geometric distortion data of texture defects. The purpose of this analysis is to identify the deformation of the surface texture in the geometric structure, such as warping, deformation or fracture. By performing a detailed evaluation of these geometric distortions, the nature and severity of the texture defects can be determined. In addition, the roughness perception evaluation provides roughness perception data on texture defects by analyzing the impact of texture defects on touch and visual effects. This data is crucial for evaluating the overall quality and user experience of the product, and helps to effectively control quality during the design and manufacturing stages. Based on the texture defect roughness perception data and geometric distortion data, the texture structure reflection map of the mobile power bank is classified into defect severity levels. This step gives a defect severity rating by comprehensively considering the roughness and geometric distortion of the texture defect. This classification report not only provides a detailed description of each defect, but also helps to formulate corresponding quality control measures. Through this systematic classification, defects of different levels can be efficiently identified and handled, thereby optimizing the production process and improving the quality of the final product, ensuring that the appearance and function of each mobile power bank meet the expected standards. Based on the defect severity classification report, the design of automated quality monitoring firmware is carried out to achieve real-time appearance quality detection of mobile power banks. By uploading the inspection data to the cloud platform, the firmware can perform real-time analysis and monitoring to ensure the automation and intelligence of quality control. The designed firmware closely combines the automated processing process with quality inspection, improving the inspection efficiency and accuracy of the production line. In addition, through the real-time data feedback of the cloud platform, it can quickly respond to quality problems in the production process, reduce the need for human intervention, and improve the overall efficiency of the production line and the consistency of the product.Therefore, the present invention optimizes a traditional method for detecting the appearance quality of a mobile power source based on machine vision, solves the problem that the manual inspection method is not only time-consuming and labor-intensive, but also easily affected by human factors, leading to inconsistent detection or missed defects, improves efficiency, reduces the impact of human factors, and reduces the problem of inconsistent detection or missed defects.

[0011] Preferably, step S1 comprises the following steps:

[0012] Step S11: using a high-resolution spectral camera to collect multi-angle spectral images of the mobile power supply to obtain a multi-angle spectral image of the mobile power supply;

[0013] Step S12: performing image contrast optimization processing on the multi-angle spectrum image of the mobile power source to obtain a spectrum optimized image of the mobile power source;

[0014] Step S13: performing image sharpening processing on the mobile power source spectrum optimization image to obtain a mobile power source spectrum sharpened image.

[0015] The present invention uses a high-resolution spectral camera to collect multi-angle spectral images of a mobile power supply to obtain multi-angle spectral images of the mobile power supply. The collection of these images captures the reflection characteristics of the surface of the mobile power supply in different spectral ranges at different angles. This multi-angle collection method can comprehensively record the texture and reflection characteristics of the surface of the object, avoiding the omission of information caused by a single perspective. The high-resolution spectral camera ensures that the details of the image are clear and can capture tiny surface irregularities. This stage lays a solid foundation for subsequent image processing and analysis. Through multi-angle data collection, the real surface condition of the mobile power supply can be more accurately reflected, and the reliability and accuracy of subsequent analysis results are improved. The multi-angle spectral image of the mobile power supply is subjected to image contrast optimization processing to obtain a spectral optimized image of the mobile power supply. The main purpose of image contrast optimization is to enhance the visibility of each detail in the image, and to make the surface features more obvious by adjusting the brightness and contrast of the image. This processing can effectively improve the visual effect of the image, making the texture details, defects and surface features more clearly visible. The optimized image can better highlight the tiny defects and detail changes on the surface, providing more accurate basic data for subsequent image sharpening processing and texture analysis. Through contrast optimization, key features in the image are magnified, which helps to improve the accuracy and effectiveness of the subsequent analysis stage. The spectral image that has been processed by contrast optimization is sharpened to obtain a spectral sharpened image of the mobile power supply. The purpose of image sharpening is to further enhance the edges and details in the image, making the surface texture and tiny defects more prominent. Sharpening emphasizes the contrast changes between pixels in the image through an algorithm, making the boundaries of the image clearer. This process not only improves the clarity of image details, but also enhances the visualization of textures and defects, making subsequent texture structure analysis and defect detection more accurate. Through spectral sharpening, defects or irregularities on the surface of the mobile power supply can be more easily identified and located, providing accurate data support for subsequent quality inspection and analysis, and ensuring the effectiveness of overall quality monitoring.

[0016] Preferably, step S2 comprises the following steps:

[0017] Step S21: using a preset mobile power spectrum standard manual to perform spectrum deviation analysis on the mobile power spectrum sharpening image to obtain mobile power spectrum deviation data;

[0018] Step S22: performing a texture structure reflection spectrum analysis on the mobile power supply spectrum sharpening image according to the mobile power supply spectrum deviation data to obtain a mobile power supply texture structure reflection spectrum;

[0019] Step S23: identifying connection interaction points of the mobile power supply texture structure reflection map to obtain texture connection interaction point data;

[0020] Step S24: performing texture ductility analysis on the mobile power supply texture structure reflection map according to the texture connection interaction point data to obtain texture connection ductility data.

[0021] The present invention uses a preset mobile power spectrum standard manual to perform spectral deviation analysis on the spectral sharpening image, thereby obtaining the spectral deviation data of the mobile power. This process can identify the spectral deviation in the image by comparing the actual spectral characteristics in the image with the spectral standards in the standard manual. These deviations include color distortion, uneven reflection and other phenomena. Analyzing these deviations can reveal potential image quality problems or surface defects. By obtaining spectral deviation data, a more accurate reference basis can be provided for subsequent texture structure analysis, ensuring an accurate understanding of the actual surface state in the subsequent analysis process, and helping to improve the reliability of the overall detection system. According to the mobile power spectrum deviation data, the spectral sharpening image is subjected to texture structure reflection spectrum analysis to obtain the texture structure reflection spectrum of the mobile power. This step reveals the structural characteristics of the surface texture, including the distribution pattern of the texture, the reflection intensity and other information, through a detailed analysis of the spectral sharpening image. This reflection spectrum analysis can provide in-depth data on the surface texture, helping to understand the performance and characteristics of the texture under different spectral conditions. Through this method, the details of the texture structure can be fully grasped, laying the foundation for subsequent texture connection analysis, thereby improving the accuracy and comprehensiveness of texture defect detection. The texture structure reflection map of the mobile power supply is identified by connecting interaction points to obtain texture connection interaction point data. The main purpose of this process is to identify the interaction points of each texture area in the texture map, which are the key positions of texture connection. By identifying the interaction points, the consistency and connection of the texture can be accurately grasped, providing specific data support for further analysis of the texture. Identifying texture connection interaction points helps to evaluate the integrity of the texture and its performance in practical applications, and provides necessary information for subsequent texture ductility analysis, making the overall quality inspection more accurate. According to the texture connection interaction point data, the texture ductility analysis of the texture structure reflection map of the mobile power supply is performed to obtain texture connection ductility data. This step evaluates the ductility and consistency of the texture in different areas by analyzing the data of texture connection interaction points. This analysis can reveal the durability and stability of the texture in practical applications, for example, whether the texture is broken or falling off. Through the texture connection ductility data, the quality of the surface texture of the mobile power supply can be further judged to ensure that it will not have serious defects or inconsistencies during use. These data play an important guiding role in optimizing the production process and improving product quality.

[0022] Preferably, step S24 comprises the following steps:

[0023] Step S241: Calculating the texture interlacing density of different connection interaction points on the mobile power supply texture structure reflection map according to the texture connection interaction point data to obtain the connection interaction point texture interlacing density data;

[0024] Step S242: performing texture morphology expansion analysis of different connection interaction points on the mobile power supply texture structure reflection map according to the connection interaction point texture interlacing density data to obtain texture morphology expansion data;

[0025] Step S243: performing high-frequency deformation interference effect analysis on the texture morphology expansion data to obtain high-frequency deformation interference effect data;

[0026] Step S244: performing texture direction stretching analysis on the texture morphology expansion data according to the high-frequency deformation interference effect data to obtain texture direction stretching data;

[0027] Step S245: performing texture ductility analysis on the reflection spectrum of the mobile power supply texture structure according to the texture direction stretching data and the high-frequency deformation interference effect data to obtain texture connection ductility data.

[0028] The present invention calculates the texture interlacing density of different connection interaction points on the texture structure reflection map of the mobile power supply according to the texture connection interaction point data, and obtains the texture interlacing density data of the connection interaction point. This process can quantify the density of the texture at different interaction points by calculating the texture interlacing density. This calculation provides the distribution of textures in different areas, so that the complexity and density of the texture can be evaluated. High interlacing density usually means that the details of the texture are more complex, while low density indicates that the texture is relatively simple. Through this step, dense and sparse areas in the texture can be identified, providing basic data for subsequent texture morphological analysis, which helps to fully understand the spatial distribution characteristics and structural complexity of the texture. According to the texture interlacing density data of the connection interaction point, the texture structure reflection map of the mobile power supply is subjected to texture morphological expansion analysis, thereby obtaining texture morphological expansion data. This step evaluates the expansion and expansion of the texture by analyzing the morphological changes of the texture at different interaction points. Texture morphological expansion analysis can reveal the expansion trend and morphological changes of the texture on the surface of the object, which is crucial for understanding the stability and durability of the texture in practical applications. Through this analysis, the potential deformation area of ​​the texture can be identified, which helps to determine whether the texture has irregular expansion or morphological distortion problems, thereby providing an important basis for subsequent quality control and optimized design. The high-frequency deformation interference effect analysis is performed on the texture morphological expansion data to obtain high-frequency deformation interference effect data. The purpose of this process is to identify the high-frequency deformation effects that occur during the expansion of the texture morphology, which include subtle texture fluctuations and deformations. High-frequency deformation interference effect analysis can detect the tiny deformations and interference phenomena in the texture, and these details are important for understanding the performance and stability of the texture under different conditions. By analyzing these effects, we can gain an in-depth understanding of the potential problems of the texture in actual use, and make targeted improvements in the design and manufacturing process to improve the overall quality and performance of the product. According to the high-frequency deformation interference effect data, the texture morphological expansion data is subjected to texture direction stretching analysis to obtain texture direction stretching data. The goal of this step is to analyze the stretching effect of the texture in different directions and evaluate the performance of the texture under the influence of stress or deformation. Texture direction stretching analysis can reveal the ductility and deformation trend of the texture in different directions, so as to determine whether the texture will have directional uneven stretching or deformation in actual use. Through this analysis, structural problems that occur during the application of textures can be identified, helping to improve texture design and optimize material properties, ensuring the consistency and stability of textures in actual use. Based on the texture directional stretching data and high-frequency deformation interference effect data, the texture ductility analysis of the texture structure reflection map of the mobile power bank is performed to obtain the texture connection ductility data. This process combines the data from the previous steps to comprehensively evaluate the ductility and connection performance of the texture. Texture ductility analysis can comprehensively consider the density, morphological expansion, high-frequency deformation and directional stretching of the texture to provide a comprehensive texture performance evaluation.Through this analysis, the overall stability and durability of the texture in actual application can be clarified, and defects and inconsistencies can be identified, providing strong data support for optimizing design and improving production, ensuring the high quality and reliability of the final product.

[0029] Preferably, step S3 comprises the following steps:

[0030] Step S31: performing texture extension and disconnection identification on the mobile power supply texture structure reflection map according to the texture connection ductility data to obtain texture extension and disconnection data;

[0031] Step S32: performing defect depth identification on the texture extension disconnection data to obtain texture extension defect depth data;

[0032] Step S33: performing geometric distortion analysis on the texture extension disconnection data according to the texture extension defect depth data to obtain texture defect geometric distortion data;

[0033] Step S34: performing roughness perception evaluation on the texture defect geometric distortion data to obtain texture defect roughness perception data.

[0034] The present invention performs texture extension and disconnection identification on the texture structure reflection map of the mobile power supply according to the texture connection ductility data, thereby obtaining texture extension and disconnection data. The main purpose of this process is to identify the extension fracture or disconnection phenomenon in the texture, that is, whether the texture is broken or separated during the expansion process. By analyzing the texture connection ductility, it is possible to accurately identify the poor connection or extension fracture areas in the texture, which usually lead to unstable texture performance or surface defects. The obtained texture extension and disconnection data can help further understand the defect characteristics of the texture in practical applications, and lay the foundation for subsequent defect depth identification and geometric distortion analysis, thereby ensuring the reliability and quality of the final product. Defect depth identification is performed on the texture extension and disconnection data to obtain texture extension defect depth data. This step determines the specific extent of the defect in the texture structure by evaluating the depth of the texture disconnection phenomenon. Depth identification can provide detailed information about disconnection defects, including the depth and range of the defect and its impact on the overall texture performance. By accurately measuring the depth of the defect, it is possible to identify which areas have more serious texture damage, thereby evaluating the potential impact of these defects on product use. This provides detailed defect feature data for further geometric distortion analysis, which helps to formulate repair measures and optimize design to improve product quality. Geometric distortion analysis is performed on the texture extension disconnection data according to the texture extension defect depth data to obtain texture defect geometric distortion data. This process evaluates the geometric deformation caused by disconnection or extension defects by analyzing the geometric characteristics of texture defects. Geometric distortion analysis can reveal the morphological changes of textures in defect areas, including dimensional deformation, irregular shape and other phenomena. This analysis helps to understand the impact of texture defects on the overall surface geometry and provides basic data for subsequent roughness perception evaluation. By analyzing the geometric distortion data, the specific deformation area of ​​the texture can be identified and corresponding repair measures can be formulated to optimize the appearance and performance of the product. Rough perception evaluation is performed on the texture defect geometric distortion data to obtain texture defect roughness perception data. This step aims to evaluate the roughness and perception effect of the texture defect area, and judge the impact of the defect on the surface roughness perception by analyzing the geometric distortion data. Rough perception evaluation can reveal the impact of texture defects on the user's touch and visual perception, and help determine the severity of the defect and its impact on the user experience. This kind of evaluation is crucial for optimizing product design and production process, because it is directly related to the appearance quality and comfort of the product. Through the roughness perception data obtained, the texture design can be improved in a targeted manner, thus improving the overall quality of the final product and user satisfaction.

[0035] Preferably, step S33 includes the following steps:

[0036] Step S331: performing distribution dispersion analysis on the texture extension defect depth data to obtain defect distribution dispersion data;

[0037] Step S332: Calculate the fault curvature according to the defect distribution dispersion data and the texture extension defect depth data to obtain texture fault curvature data;

[0038] Step S333: performing geometric cluster analysis on the texture fault curvature data to obtain texture fault geometric data;

[0039] Step S334: performing geometric distortion analysis on the texture extension and disconnection data according to the texture fault geometric morphology data to obtain texture defect geometric distortion data.

[0040] The present invention performs distribution discreteness analysis on texture extension defect depth data to obtain defect distribution discreteness data. This process evaluates the distribution of defects in the texture structure by calculating the discreteness of the texture defect depth. Discreteness analysis can reveal whether the spatial distribution of defects is uniform and whether the severity of defects is concentrated in a specific area. This analysis helps to understand the extensiveness and local concentration of defects in the texture during extension, and provides basic data for subsequent fault curvature calculation and geometric morphology analysis. Through discreteness data, key defect areas in the texture can be identified, which helps to more accurately formulate repair strategies and optimize designs. Fault curvature calculation is performed based on defect distribution discreteness data and texture extension defect depth data to obtain texture fault curvature data. The purpose of this step is to calculate the curvature of the fault in the texture by analyzing the distribution and depth data of the defects. This curvature calculation can reveal the degree of geometric change in the texture defect area and provide detailed information on how the fault is bent or twisted. Fault curvature data helps to identify complex deformation conditions in the texture, such as bending, bulging or depression, thereby providing specific geometric feature data for subsequent geometric morphology analysis. This calculation can help designers understand the impact of defects on the texture surface, thereby optimizing the structural design and performance of the texture. The texture fault curvature data is subjected to geometric clustering analysis to obtain the texture fault geometric data. The goal of this step is to classify the curvature data of the fault in the texture to identify defect areas with similar geometric characteristics. Geometric clustering analysis can group different types of fault curvatures and reveal the common characteristics and patterns of texture defects. This analysis can help determine the main types of texture defects and their geometric characteristics, thereby providing valuable information for texture optimization and improvement. By identifying different geometric data, more effective repair strategies can be formulated to improve the overall quality and function of the texture. According to the texture fault geometric data, the texture extension disconnection data is subjected to geometric distortion analysis to obtain the texture defect geometric distortion data. This process analyzes the overall geometric deformation of the texture by combining the geometric data of the fault. Geometric distortion analysis can reveal the specific morphological changes of the texture in the defect area, including the degree and pattern of distortion. This analysis helps to understand how defects affect the geometric stability and consistency of the texture, thereby providing a basis for subsequent design improvements and performance improvements. Through geometric distortion data, structural problems in texture can be identified, and effective measures can be taken to repair and optimize them, ensuring the high quality and reliability of the product.

[0041] Preferably, performing geometric cluster analysis on texture fault curvature data comprises the following steps:

[0042] Calculating the boundary difference of the texture fault curvature data to obtain the texture fault curvature boundary difference data;

[0043] According to the texture fault curvature boundary difference data, the texture fault curvature data is analyzed for morphological trend, and the texture fault morphological trend data is obtained;

[0044] According to the texture fault curvature boundary difference data and texture fault morphology trend data, geometric morphology cluster analysis is performed to obtain texture fault geometric morphology data.

[0045] The present invention calculates the boundary difference of texture fault curvature data to obtain texture fault curvature boundary difference data. This step can evaluate the change of fault curvature in different regions by calculating the boundary difference of texture fault curvature data. The boundary difference calculation reveals the degree of variation of texture fault boundaries and the influence of these variations on the overall texture structure. This data helps to identify the areas with obvious curvature changes in the texture, so that the problem areas can be located more accurately, providing basic information for subsequent morphological trend analysis and geometric morphological clustering. By understanding the boundary difference, designers can better understand the geometric characteristics and deformation patterns of the fault, thereby providing an effective basis for texture optimization and defect repair. According to the texture fault curvature boundary difference data, the texture fault curvature data is subjected to morphological trend analysis to obtain texture fault morphological trend data. This step uses the boundary difference data to analyze the morphological trend of the texture fault and evaluate the trend and directionality of the texture defect. The morphological trend analysis can reveal the changing trend of the fault curvature, such as the extension direction of the defect and the changing pattern of the curvature. This analysis provides geometric trend information of texture faults, which helps to understand the distribution law and morphological characteristics of faults, thus providing valuable reference when processing and optimizing texture structures. Through this analysis, the main directions and morphological patterns of texture defects can be more accurately identified, so as to formulate more effective repair strategies. Geometric morphological clustering analysis is performed based on the texture fault curvature boundary difference data and the texture fault morphological trend data to obtain the texture fault geometric morphological data. This process combines the boundary difference and morphological trend data, and classifies the fault data with similar geometric characteristics through cluster analysis to obtain the texture fault geometric morphological data. Geometric morphological clustering analysis can reveal the different types and distribution patterns of faults in textures, and provide a detailed geometric description for each type of fault. This analysis helps to identify the key fault types and their distribution characteristics in textures, so that the texture structure can be optimized and improved in a targeted manner. Through geometric morphological data, the defect characteristics in textures can be more comprehensively understood, thereby improving the design accuracy of textures and product quality.

[0046] Preferably, step S34 includes the following steps:

[0047] Step S341: performing a concave-convex distribution structure analysis on the texture defect geometric distortion data based on the texture extension defect depth data to obtain the texture defect concave-convex distribution data;

[0048] Step S342: performing contact force simulation on the texture defect concave-convex distribution data to obtain the concave-convex distribution contact force data;

[0049] Step S343: calculating the holding friction coefficient for the texture defect concave-convex distribution data according to the concave-convex distribution contact force data to obtain the texture defect holding friction coefficient;

[0050] Step S344: performing roughness perception evaluation on the texture defect geometric distortion data according to the texture defect grip friction coefficient and the concave-convex distribution contact force data to obtain texture defect roughness perception data.

[0051] The present invention performs a concave-convex distribution structure analysis on the texture defect geometric distortion data based on the texture extension defect depth data to obtain the texture defect concave-convex distribution data. This step identifies and depicts the concave-convex distribution structure of the texture defect area by analyzing the texture extension defect depth data. This process can reveal the surface contour of the texture defect in detail, including the specific position and degree of the depression and protrusion. Through the analysis of the concave-convex distribution structure, the geometric feature data of the texture surface can be obtained, which is crucial for understanding the impact of defects on the overall structure and function of the texture. The concave-convex distribution data can help designers identify the main defect areas and their geometric shapes in the texture, thereby laying the foundation for subsequent contact force simulation and roughness perception evaluation. The contact force simulation is performed on the texture defect concave-convex distribution data to obtain the concave-convex distribution contact force data. This step can calculate the distribution of forces received by different concave-convex areas in actual applications by performing a contact force simulation on the concave-convex distribution data of the texture defect. The simulation process takes into account the response of different concave-convex structures when subjected to force, and can provide information about the mechanical behavior of the defect area under actual use conditions. Contact force data helps understand how texture defects affect surface force transmission and its performance in actual operation. This analysis can not only evaluate the strength and durability of the texture, but also provide a scientific basis for optimizing texture design and improving product performance. The grip friction coefficient of the texture defect concave-convex distribution data is calculated based on the concave-convex distribution contact force data to obtain the texture defect grip friction coefficient. This step uses the results of the contact force simulation to calculate the grip friction coefficient of the texture surface. The grip friction coefficient is a key indicator for evaluating the friction performance of the texture surface during contact and operation. By calculating the grip friction coefficient, the effect of texture defects on friction can be quantified to understand the grip or friction performance in actual use. This data can help designers optimize the structure of the texture surface to improve the operating comfort and safety of the product, especially when it involves hand operation or contact with the surface. The texture defect geometric distortion data is evaluated for roughness perception based on the texture defect grip friction coefficient and the concave-convex distribution contact force data to obtain the texture defect roughness perception data. This step evaluates the roughness perception of the geometric distortion of the texture defect by combining the grip friction coefficient and the contact force data. Roughness perception data reflects the tactile and visual perception of textured surfaces in actual use, including the subjective perception of surface roughness. By evaluating roughness perception data, we can understand how texture defects affect the user's perception of textured surfaces and help optimize the design to improve the user experience. This evaluation is important for determining the tactile quality and comfort of the product, which can guide design improvements to achieve better product performance and user satisfaction.

[0052] Preferably, step S4 comprises the following steps:

[0053] Step S41: constructing a defect anomaly map for the mobile power supply texture structure reflection map according to the texture defect roughness perception data and the texture defect geometric distortion data to obtain a texture structure defect anomaly map;

[0054] Step S42: performing feature dimensionality reduction on the texture structure defect anomaly map to obtain a texture structure defect anomaly dimensionality reduction map;

[0055] Step S43: Based on the generative adversarial network, the texture structure defect anomaly dimensionality reduction map is classified into defect severity levels to obtain a defect severity level classification report.

[0056] The present invention constructs a defect anomaly map for the mobile power supply texture structure reflection map according to the texture defect roughness perception data and the texture defect geometric distortion data to obtain a texture structure defect anomaly map. This step constructs a texture structure defect anomaly map by combining the roughness perception data and the geometric distortion data. The map shows the abnormal area on the texture surface and can visualize the type, distribution and severity of the defect. This map provides designers with detailed texture defect information, helps to identify and locate potential problems in the texture, and is particularly important for product quality control. The construction of the defect anomaly map can improve the accuracy of defect detection, so that further analysis and improvement measures can be more targeted and effective. The texture structure defect anomaly map is subjected to feature dimensionality reduction to obtain a texture structure defect anomaly dimensionality reduction map. This step uses feature dimensionality reduction technology to simplify the map data, reduce the complexity of the data, and retain key feature information. Feature dimensionality reduction helps to reduce computational complexity, improve the efficiency of subsequent analysis, and make data visualization and interpretation more intuitive by reducing data dimensions. The reduced-dimensional map is easier to perform data mining and pattern recognition, thereby revealing the potential patterns and trends of texture defects. This technology can enhance data processing capabilities, make the analysis of defect anomaly maps more efficient, and support further classification and evaluation work. Based on the generative adversarial network, the defect severity level of the texture structure defect anomaly reduced dimension map is classified to obtain a defect severity classification report. This step uses the generative adversarial network (GAN) to classify the defect map after dimensionality reduction to evaluate the severity of the defects. By learning a large amount of sample data, the generative adversarial network can effectively identify defect features of different levels and accurately classify them. The classification report provides the severity of each defect, ranging from minor to severe, to help designers prioritize the most critical defect issues. Through this advanced analysis technology, the quality of the texture structure can be more comprehensively evaluated, the production process can be optimized, and the overall performance and reliability of the final product can be improved.

[0057] Preferably, step S5 comprises the following steps:

[0058] Step S51: normalizing the defect severity classification report to obtain a defect severity classification normalized report;

[0059] Step S51: Designing an automated quality monitoring firmware based on the defect severity classification normalization report to obtain an automated quality monitoring firmware;

[0060] Step S52: Send the automated quality monitoring firmware to the cloud platform to perform real-time appearance quality detection.

[0061] The present invention normalizes the defect severity classification report to obtain a defect severity classification normalization report. In this step, defect data of different ranges and scales are converted into a unified standard format by normalizing the defect severity classification report. Normalization eliminates the influence of data dimension, so that the defect severity information remains consistent between different data sets and analysis stages. This processing method makes the classification result more reliable and facilitates cross-system data comparison and comprehensive analysis. The normalized report provides standardized input data for the subsequent automated quality monitoring system, which helps to improve the accuracy and stability of the system, thereby ensuring the consistency and effectiveness of the quality monitoring process. Based on the defect severity classification normalization report, an automated quality monitoring firmware is designed to obtain an automated quality monitoring firmware. In this step, a special firmware is designed according to the normalized report to realize automated quality monitoring. The automated quality monitoring firmware integrates the function of processing defect data and can detect and analyze texture defects in the production process in real time. The designed firmware will have efficient data processing capabilities and response mechanisms, can automatically perform quality inspection tasks, reduce manual intervention, and improve monitoring efficiency. By applying normalized defect information to firmware design, it is possible to ensure that the monitoring system can accurately identify various defect levels and dynamically adjust and optimize the production process, thereby improving overall product quality and production efficiency. Send the automated quality monitoring firmware to the cloud platform to perform real-time appearance quality detection. This step involves deploying the developed automated quality monitoring firmware to the cloud platform to achieve real-time monitoring of the production process. The advantage of the cloud platform lies in its powerful computing power and extensive data storage capabilities, making real-time detection and data analysis more efficient and flexible. By uploading the firmware to the cloud platform, remote monitoring and instant data feedback across locations can be achieved, supporting comprehensive coverage of large-scale production lines. The real-time detection function can detect and report quality problems in a timely manner, allowing the production line to respond quickly and make adjustments. This not only improves the transparency of the production process, but also enhances the timeliness and accuracy of quality management, thereby optimizing the appearance quality and overall performance of the product.

[0062] The beneficial effect of the present invention is that a spectral sharpening image of a mobile power supply can be obtained by collecting multi-angle spectral images of the mobile power supply through a high-resolution spectral camera. These images reveal the reflection characteristics of the mobile power supply at different wavelengths, so that the subtle differences on the surface of the object can be captured more clearly. This spectral sharpening not only improves the clarity and contrast of the image, but also emphasizes subtle surface features, laying the foundation for subsequent texture analysis. By collecting data from multiple angles, a comprehensive understanding of the surface state of the mobile power supply can be ensured to avoid potential defects missed due to a single viewing angle. The texture structure reflection spectrum analysis of the acquired mobile power supply spectrum is performed to obtain the texture structure reflection spectrum of the mobile power supply. This analysis process mainly focuses on identifying and parsing the texture pattern of the mobile power supply surface, including its reflection characteristics and the distribution of surface texture. Through the analysis of the texture structure reflection spectrum, the details of the surface texture and its ductility can be deeply understood. These data are further used for texture ductility analysis to help evaluate the behavior and coherence of the surface texture under different conditions, thereby providing a comprehensive understanding of the texture and laying a solid foundation for subsequent defect detection and quality control. The geometric distortion analysis of texture extension and disconnection data is performed to obtain the geometric distortion data of texture defects. The purpose of this analysis is to identify the deformation of the surface texture in the geometric structure, such as warping, deformation or fracture. By performing a detailed evaluation of these geometric distortions, the nature and severity of the texture defects can be determined. In addition, the roughness perception evaluation provides roughness perception data on texture defects by analyzing the impact of texture defects on touch and visual effects. This data is crucial for evaluating the overall quality and user experience of the product, and helps to effectively control quality during the design and manufacturing stages. Based on the texture defect roughness perception data and geometric distortion data, the texture structure reflection map of the mobile power bank is classified into defect severity levels. This step gives a defect severity rating by comprehensively considering the roughness and geometric distortion of the texture defect. This classification report not only provides a detailed description of each defect, but also helps to formulate corresponding quality control measures. Through this systematic classification, defects of different levels can be efficiently identified and handled, thereby optimizing the production process and improving the quality of the final product, ensuring that the appearance and function of each mobile power bank meet the expected standards. Based on the defect severity classification report, the design of automated quality monitoring firmware is carried out to achieve real-time appearance quality detection of mobile power banks. By uploading the inspection data to the cloud platform, the firmware can perform real-time analysis and monitoring to ensure the automation and intelligence of quality control. The designed firmware closely combines the automated processing process with quality inspection, improving the inspection efficiency and accuracy of the production line. In addition, through the real-time data feedback of the cloud platform, it can quickly respond to quality problems in the production process, reduce the need for human intervention, and improve the overall efficiency of the production line and the consistency of the product.Therefore, the present invention optimizes a traditional method for detecting the appearance quality of a mobile power source based on machine vision, solves the problem that the manual inspection method is not only time-consuming and labor-intensive, but also easily affected by human factors, leading to inconsistent detection or missed defects, improves efficiency, reduces the impact of human factors, and reduces the problem of inconsistent detection or missed defects. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 The figure is a flow chart of the steps of a method for detecting the appearance quality of a mobile power source based on machine vision;

[0064] Figure 2 for Figure 1 Detailed implementation steps of step S2 in FIG.

[0065] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0066] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0067] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0068] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0069] To achieve this, please refer to Figure 1 to Figure 2, a method for detecting the appearance quality of a mobile power source based on machine vision, the method comprising the following steps:

[0070] Step S1: using a high-resolution spectral camera to collect multi-angle spectral images of the mobile power supply to obtain a spectral sharpened image of the mobile power supply;

[0071] Step S2: performing texture structure reflection spectrum analysis on the mobile power source spectral sharpening image to obtain the mobile power source texture structure reflection spectrum; performing texture ductility analysis on the mobile power source texture structure reflection spectrum to obtain texture connection ductility data;

[0072] Step S3: performing geometric distortion analysis on the texture extension disconnection data to obtain texture defect geometric distortion data; performing roughness perception evaluation on the texture defect geometric distortion data to obtain texture defect roughness perception data;

[0073] Step S4: classifying the defect severity level of the mobile power supply texture structure reflection map according to the texture defect roughness perception data and the texture defect geometric distortion data to obtain a defect severity level classification report;

[0074] Step S5: Design an automated quality monitoring firmware based on the defect severity classification report to obtain the automated quality monitoring firmware, and send the automated quality monitoring firmware to the cloud platform to perform real-time appearance quality detection.

[0075] In the embodiment of the present invention, reference Figure 1 The above is a schematic flow chart of the steps of a mobile power supply appearance quality detection method based on machine vision of the present invention. In this example, the mobile power supply appearance quality detection method based on machine vision includes the following steps:

[0076] Step S1: using a high-resolution spectral camera to collect multi-angle spectral images of the mobile power supply to obtain a spectral sharpened image of the mobile power supply;

[0077] In an embodiment of the present invention, a high-resolution spectral camera is installed, and it is ensured to have a multi-angle capture function. The mobile power supply is placed on a rotatable platform, and the platform needs to be precisely rotated so that the camera can collect comprehensive spectral images of the surface of the mobile power supply from multiple angles (e.g., 0°, 45°, 90°, 135°, etc.). During the acquisition process, the light source should remain constant and uniform to avoid data errors caused by inconsistent brightness. The image captured by the camera is subjected to real-time spectral sharpening processing by a spectral analysis unit to eliminate the interference of surface reflection and obtain a high-precision spectral sharpened image of the mobile power supply. In the sharpened image, the surface texture, color, reflection characteristics, etc. are fully reflected, providing a basis for subsequent analysis.

[0078] Step S2: performing texture structure reflection spectrum analysis on the mobile power source spectral sharpening image to obtain the mobile power source texture structure reflection spectrum; performing texture ductility analysis on the mobile power source texture structure reflection spectrum to obtain texture connection ductility data;

[0079] In an embodiment of the present invention, the spectral sharpening image of the mobile power supply is analyzed, and each pixel in the image represents a specific spectral reflection characteristic. By using a high-precision image processing algorithm, the spectral reflection intensity and angle information of each band are extracted pixel by pixel, thereby generating a texture structure reflection spectrum of the mobile power supply. The spectrum records the spectral reflection characteristics of each area by segmenting different areas in the image. Next, an in-depth texture ductility analysis is performed on the texture structure reflection spectrum. The specific steps are to evaluate the inter-pixel connectivity of each texture area, calculate the degree of extension between texture nodes, and record the continuity and fracture of the texture. This analysis pays special attention to detecting the ductility of the texture at different angles, and determines the actual connection of the texture by comparing the reflection differences of the texture in multiple spectral bands. The texture connection ductility data finally output will record the integrity of the texture, the location of the breakpoint, and the ductility coefficient of each texture area, laying the foundation for subsequent analysis.

[0080] Step S3: performing geometric distortion analysis on the texture extension disconnection data to obtain texture defect geometric distortion data; performing roughness perception evaluation on the texture defect geometric distortion data to obtain texture defect roughness perception data;

[0081] In an embodiment of the present invention, the texture connection ductility data generated in step S2 is processed. By using a geometric deformation analysis tool, the specific morphological changes of the extended disconnection area are identified. The geometric shape distortion of the fracture area is analyzed in detail, including the offset of the extension direction, the amplitude of the texture distortion, etc. In the analysis, a precise geometric algorithm is used to calculate the shape parameters of each fracture area, such as the fracture angle, deformation rate, and spatial distortion, and finally the texture defect geometric distortion data is formed. This data describes the geometric distortion of the surface texture during extension and provides a basis for the in-depth perception of surface defects. Next, the texture defect geometric distortion data is used for rough perception evaluation. During the evaluation process, the microscopic roughness of the texture defect area is detected by a detailed analysis of the surface reflected light. The specific perception method is to use high-resolution spectral reflectance data to measure the microscopic differences in light scattering and absorption on the texture surface. By comparing the roughness coefficients of different areas, the severity of the surface defects is accurately judged, and finally the texture defect rough perception data is generated. This data reflects the perceived roughness of the texture defect and is used for further surface quality evaluation.

[0082] Step S4: classifying the defect severity level of the mobile power supply texture structure reflection map according to the texture defect roughness perception data and the texture defect geometric distortion data to obtain a defect severity level classification report;

[0083] In the embodiment of the present invention, based on the texture defect roughness perception data and texture defect geometric distortion data generated in step S3, the texture structure reflection map of the mobile power supply is classified into defect severity levels. The specific operation includes a joint analysis of the degree of geometric distortion and roughness of each defect area. In the classification process, a multi-dimensional data standard is adopted. First, the defects are preliminarily divided into slight, medium, severe and other levels according to the parameters such as the distortion angle, shape offset and distortion rate in the geometric distortion data; then, the classification is further refined by combining the surface roughness difference in the roughness perception data. For example, a larger distortion area and a significant roughness area are classified as a "severe defect", while a smaller distortion area and an insignificant roughness area are classified as a "slight defect". In the classification process, a specific physical threshold standard is adopted (for example, an area with a geometric distortion exceeding 0.5° is classified as a "medium defect") to ensure that the evaluation results of each defect are accurate and consistent. After the classification is completed, a defect severity classification report is generated, which lists the specific location, severity level and detailed physical parameters of each defect, providing an accurate basis for the subsequent quality monitoring firmware design.

[0084] Step S5: Design an automated quality monitoring firmware based on the defect severity classification report to obtain the automated quality monitoring firmware, and send the automated quality monitoring firmware to the cloud platform to perform real-time appearance quality detection.

[0085] In an embodiment of the present invention, based on the defect severity classification report generated in step S4, the design of the automated quality monitoring firmware is carried out. First, according to the defect level classified in the report, specific monitoring rules and detection standards are set. Each defect level corresponds to a different detection threshold, for example, "serious defects" require real-time alarm processing, while "minor defects" only require routine monitoring. During the design process, the firmware includes a comprehensive detection algorithm for texture reflection, geometric distortion and surface roughness. The core algorithm of the firmware uses a preset physical threshold standard to quickly analyze and judge the image data collected during the detection process. After the design is completed, the firmware is integrated into the quality inspection equipment of the production line to achieve fully automated real-time monitoring of the appearance of the mobile power supply. To ensure the efficient execution of the detection process, the firmware sends the detection results to the cloud platform for storage and analysis via wireless transmission. The cloud platform has the ability to continuously update, and dynamically adjusts and optimizes according to the quality data collected during the production process, thereby achieving real-time monitoring and intelligent analysis of the appearance quality.

[0086] Preferably, step S1 comprises the following steps:

[0087] Step S11: using a high-resolution spectral camera to collect multi-angle spectral images of the mobile power supply to obtain a multi-angle spectral image of the mobile power supply;

[0088] Step S12: performing image contrast optimization processing on the multi-angle spectrum image of the mobile power source to obtain a spectrum optimized image of the mobile power source;

[0089] Step S13: performing image sharpening processing on the mobile power source spectrum optimization image to obtain a mobile power source spectrum sharpened image.

[0090] In an embodiment of the present invention, a high-resolution spectral camera is used to perform multi-angle spectral image acquisition operation on a mobile power supply. In order to ensure that the spectral information of the appearance of the mobile power supply can be accurately captured at different angles, the camera needs to perform image acquisition at multiple angles (such as 0°, 45°, 90°, etc.). In the specific implementation process, the mobile power supply is placed on a rotatable fixture, and the spectral camera performs multi-angle capture around the horizontal and vertical directions of the mobile power supply. Each captured image contains complete spectral information, including spectral distribution from ultraviolet to near-infrared bands. In order to ensure image quality, the optical sensor of the camera automatically adjusts the aperture and exposure time to adapt to light changes at different angles to avoid overexposure or underexposure of the image. The collected spectral image is transmitted to the data processing unit and saved in the image database, and finally a multi-angle spectral image of the mobile power supply is generated. Based on the multi-angle spectral image of the mobile power supply generated in step S11, the image contrast optimization process is performed next. First, by calculating the global and local contrast of the image, the area with uneven brightness distribution or insufficient grayscale range in the spectral image is identified. During the optimization process, histogram equalization technology is used to redistribute the brightness values ​​of pixels, thereby expanding the dynamic range of the image and making the details more clearly presented. In order to avoid the increase of image noise caused by over-enhancement, high-frequency noise needs to be filtered out through denoising. The specific steps include enhancing the low-brightness area of ​​the image to improve the details of the dark area and avoid oversaturation in the bright area. The optimized image has better contrast and can show the subtle differences between different materials and textures on the surface of the mobile power supply. Finally, the mobile power supply spectrum optimization image is obtained, which provides a good foundation for subsequent image sharpening processing. After obtaining the mobile power supply spectrum optimization image, image sharpening processing is performed to improve the edge clarity and texture details of the image. The goal of sharpening processing is to enhance the details of the mobile power supply surface, especially the texture and boundary parts. The specific method uses Gaussian difference algorithm (DoG) or Laplace sharpening technology to identify the areas in the image where the grayscale value changes significantly by calculating the second-order derivative of the image, and enhance these areas. During the processing, the neighborhood information of each pixel is compared with the surrounding pixels, and the part with larger changes is enhanced, so that the edge of the object is clearer. In order to avoid noise enhancement caused by the sharpening process, the image is post-processed using adaptive filtering technology to eliminate the introduced fine noise. During the sharpening process, special attention should be paid to the preservation and enhancement of the surface texture, so that the appearance characteristics of the mobile power bank are more prominent visually. After the processing is completed, a spectral sharpened image of the mobile power bank is generated for subsequent texture structure analysis.

[0091] Preferably, step S2 comprises the following steps:

[0092] Step S21: using a preset mobile power spectrum standard manual to perform spectral deviation analysis on the mobile power spectrum sharpening image to obtain mobile power spectrum deviation data;

[0093] Step S22: performing texture structure reflection spectrum analysis on the mobile power supply spectrum sharpening image according to the mobile power supply spectrum deviation data to obtain a mobile power supply texture structure reflection spectrum;

[0094] Step S23: identifying connection interaction points of the mobile power supply texture structure reflection map to obtain texture connection interaction point data;

[0095] Step S24: performing texture ductility analysis on the mobile power supply texture structure reflection map according to the texture connection interaction point data to obtain texture connection ductility data.

[0096] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes:

[0097] Preferably, step S2 comprises the following steps:

[0098] Step S21: using a preset mobile power spectrum standard manual to perform spectral deviation analysis on the mobile power spectrum sharpening image to obtain mobile power spectrum deviation data;

[0099] In an embodiment of the present invention, a spectral deviation analysis is performed on the mobile power supply spectrum sharpening image using a preset mobile power supply spectrum standard manual. First, it is necessary to determine the standard reflection characteristics of the mobile power supply appearance in different bands according to the various spectral features in the spectrum standard manual. The spectrum standard manual contains the reflection spectrum information of different materials, colors, and textures of the mobile power supply under standard lighting conditions as a comparison benchmark. The first step of the spectral deviation analysis is to extract the spectral information in the mobile power supply spectrum sharpening image and separate and analyze the spectral values ​​of different bands. The specific operation includes obtaining the spectral characteristic curve of each pixel through a spectral analysis device and comparing it with the standard spectral curve. During the comparison process, a differential processing technology is used to calculate the deviation between the actual collected spectrum and the standard spectrum, including parameters such as spectral intensity, peak and trough position, and reflectivity difference. Through these calculations, spectral deviation data is obtained to characterize the degree of deviation of the mobile power supply surface in different bands. The mobile power supply spectrum deviation data finally generated provides basic data support for subsequent texture structure analysis.

[0100] Step S22: performing texture structure reflection spectrum analysis on the mobile power supply spectrum sharpening image according to the mobile power supply spectrum deviation data to obtain a mobile power supply texture structure reflection spectrum;

[0101] In an embodiment of the present invention, after obtaining the mobile power supply spectrum deviation data of step S21, the texture structure reflection spectrum analysis of the mobile power supply spectrum sharpening image is carried out. First, by applying the spectrum deviation data to the spectrum sharpening image, the spectrum reflection characteristics of each image pixel are recalibrated to eliminate the influence of the spectrum deviation. Next, the reflection characteristics of the surface texture of the mobile power supply are analyzed by using image processing technology. This process extracts the spatial distribution characteristics of the surface texture of the mobile power supply by detecting the brightness gradient change and edge information in the image. Fourier transform technology is used in the analysis process to convert the texture features in the image into the frequency domain and identify the distribution law of different texture structures in spatial frequency. By comparing the energy distribution of each frequency band, the reflection intensity of different areas on the surface of the mobile power supply and its changing trend at different angles are determined. Finally, these analysis results are summarized to form a mobile power supply texture structure reflection spectrum, which characterizes the texture structure and optical reflection characteristics of the mobile power supply appearance surface in detail, and provides accurate texture basic data for further defect detection.

[0102] Step S23: identifying connection interaction points of the mobile power supply texture structure reflection map to obtain texture connection interaction point data;

[0103] In an embodiment of the present invention, after completing the construction of the reflection map of the mobile power supply texture structure, further analysis of the map is carried out to identify the texture connection interaction points. First, the texture boundary in the texture structure reflection map is accurately extracted using an edge detection algorithm, and the extracted boundary data is used to define the boundary range of each texture element in the texture structure. Next, the topological analysis method is used to structure each texture structure and analyze the geometric connection relationship between the texture structures. By calculating the geometric center of the texture structure, the boundary overlap area and the correlation between adjacent pixels, the connection points and interaction areas between different textures are identified. The recognition process pays special attention to analyzing the interaction characteristics of the texture structure in different spectral bands to ensure that the identified interaction points are not only based on a single spectral feature. Each interaction point is classified and marked according to its position in the texture structure, the change in spectral reflection intensity and the geometric morphological characteristics. Finally, the texture connection interaction point data is generated, which includes the coordinates of each interaction point, the connection strength and the description of the relationship with the adjacent texture, providing important basic data for the subsequent texture ductility analysis.

[0104] Step S24: performing texture ductility analysis on the mobile power supply texture structure reflection map according to the texture connection interaction point data to obtain texture connection ductility data.

[0105] In an embodiment of the present invention, after obtaining the texture connection interaction point data, a texture ductility analysis is performed to evaluate the continuity and integrity of the mobile power surface texture. First, a texture extension model is constructed based on the coordinates and connection strength of each interaction point. The model determines the ductility of the texture in space by calculating the connection path between the interaction points. In the specific analysis process, the minimum path algorithm is used to calculate the shortest connection path between each interaction point, and combined with the change trend of the texture reflection intensity, it is determined whether the texture has ductility characteristics. For textures with good ductility, the connection path between the interaction points is relatively smooth, and the reflection intensity changes less. Conversely, textures with poor ductility will show sudden changes in reflection intensity or breaks in the connection path. Further analysis is performed to determine whether there are breaks, jumps or overlapping areas during the texture extension process, and this information is used to determine whether the integrity and optical performance of the texture are damaged. Finally, the generated texture connection ductility data contains the degree of ductility and spatial distribution of each texture structure, providing detailed data support for subsequent geometric distortion analysis and defect detection.

[0106] Preferably, step S24 includes the following steps:

[0107] Step S241: Calculating the texture interlacing density of different connection interaction points on the mobile power supply texture structure reflection map according to the texture connection interaction point data to obtain the connection interaction point texture interlacing density data;

[0108] Step S242: performing texture morphology expansion analysis of different connection interaction points on the mobile power supply texture structure reflection map according to the connection interaction point texture interlacing density data to obtain texture morphology expansion data;

[0109] Step S243: performing high-frequency deformation interference effect analysis on the texture morphology expansion data to obtain high-frequency deformation interference effect data;

[0110] Step S244: performing texture direction stretching analysis on the texture morphology expansion data according to the high-frequency deformation interference effect data to obtain texture direction stretching data;

[0111] Step S245: performing texture ductility analysis on the reflection spectrum of the mobile power supply texture structure according to the texture direction stretching data and the high-frequency deformation interference effect data to obtain texture connection ductility data.

[0112] In an embodiment of the present invention, after obtaining the texture connection interaction point data, the texture interlacing density of different connection interaction points is calculated for the mobile power supply texture structure reflection spectrum. The calculation process is carried out by finely measuring the texture distribution around each interaction point. Specifically, a density estimation algorithm is used to analyze the texture interlacing degree around the texture interaction point based on the distance between the coordinates of the interaction point and the adjacent texture structure. The higher the density, the stronger the connectivity and better the ductility of the texture structure around the interaction point. In order to ensure the accuracy of the calculation, the texture interlacing density is compensated using multi-angle spectral reflection data to prevent errors caused by surface angle changes or light reflection. This step obtains the texture interlacing density data of each connection interaction point by point calculation, which provides a basis for the subsequent texture morphology analysis. After obtaining the texture interlacing density data of the connection interaction point, the texture structure reflection spectrum of the mobile power supply is subjected to texture morphology expansion analysis using the expansion analysis method. In the specific implementation, each interaction point is selected as the expansion center, and the expansion area is gradually expanded outward according to its interlacing density level until the expansion area overlaps with the adjacent texture boundary or reaches a preset density threshold. The expansion process not only considers the geometric morphology of the texture, but also introduces changes in the spectral reflection intensity to ensure that the expanded area is consistent with the original texture structure. For high-density areas, the expansion degree is low, while for low-density areas, the expansion range is larger. The expanded data structure is recorded as texture morphological expansion data, which is used to describe the overall expansion of the texture and its impact on the connectivity of the surface texture. On the basis of the texture morphological expansion analysis, a high-frequency deformation interference effect analysis is further performed. This analysis aims to detect the optical interference phenomenon caused by the slight deformation of the material surface during the texture expansion process. In the specific operation, high-frequency spectral data is used to perform high-frequency deformation detection on the texture of the expanded area, and the shape change of the expanded area and the slight deviation of light reflection are measured. This process uses Fourier transform technology to convert the spatial frequency information of the expanded texture into the spectrum domain, and by analyzing the changes in the high-frequency components, the deformation interference effect on the texture surface is identified. The generated high-frequency deformation interference effect data includes information such as interference amplitude, frequency distribution, and deformation intensity, which can intuitively reflect the microscopic deformation of the surface texture of the mobile power supply. According to the high-frequency deformation interference effect data, the texture morphological expansion data is subjected to texture direction stretching analysis. This analysis is based on the geometric changes of textures in different directions. By directional measurement of the texture shape in each expansion area, the trend of texture changes in the direction of tension, compression or bending is identified. First, the boundary of the texture structure is directionally extracted using the directional gradient method to calculate the main direction of each expansion area. Then, by calculating the stretch rate in the main direction, the stretch direction of the texture and its tensile strength are identified. In particular, areas with significant stretch changes are marked as potential texture defect areas. The texture direction stretch data contains the stretch direction, stretch degree, and corresponding expansion area, providing the basis for the final ductility analysis.After comprehensively analyzing the texture directional stretching data and high-frequency deformation interference effect data, the texture ductility analysis is performed. This step judges the overall ductility of the texture structure based on the comprehensive information of the stretching degree, directional distribution and interference effect in the expansion area. First, the tensile strength of each expansion area is matched with its high-frequency deformation interference effect to identify areas with high ductility and low interference effect, indicating that these areas have good texture ductility. On the contrary, areas with high interference effects and large-direction stretching are marked as parts with poor ductility. Finally, the generated texture connection ductility data comprehensively describes the ductility of the surface texture of the mobile power supply, providing a clear basis for subsequent quality inspection and defect judgment.

[0113] Preferably, step S3 comprises the following steps:

[0114] Step S31: performing texture extension and disconnection identification on the mobile power supply texture structure reflection map according to the texture connection ductility data to obtain texture extension and disconnection data;

[0115] Step S32: performing defect depth identification on the texture extension disconnection data to obtain texture extension defect depth data;

[0116] Step S33: performing geometric distortion analysis on the texture extension disconnection data according to the texture extension defect depth data to obtain texture defect geometric distortion data;

[0117] Step S34: performing roughness perception evaluation on the texture defect geometric distortion data to obtain texture defect roughness perception data.

[0118] In an embodiment of the present invention, based on the texture connection ductility data, the texture extension disconnection identification is performed on the reflection spectrum of the mobile power supply texture structure. During operation, the texture spectrum is first segmented at the pixel level by a refinement algorithm to identify the breakpoints and discontinuous areas in the texture extension process. Subsequently, the edge detection technology is used to accurately analyze the texture boundary and identify the broken area in the texture extension. This process analyzes the disconnection phenomenon caused by stretching, compression or surface defects in the texture structure during the extension process by comparing the ductility and local continuity of the texture one by one. The disconnection area usually manifests as a sudden change in texture density or incomplete connection. By comparing the texture structure morphology of different disconnection areas, texture extension disconnection data is finally generated to provide basic support for subsequent defect depth identification. Based on the texture extension disconnection data, the defect depth identification is performed for each disconnection area. In this step, the surface profile scanning technology is used to measure the depth of the texture structure in the disconnection area. During the specific operation, the depth change of the disconnection is detected by using the spectral reflection difference and the change of the texture geometric profile. By measuring the disconnection points at different positions point by point, the fracture of the texture structure at different depths can be identified. Combined with the surface optical properties of the texture structure, the relationship between the surface defects and the deep fractures is analyzed, and finally the texture extension defect depth data is generated. This data can clearly show the depth of the texture disconnection, providing a basis for judging whether the surface quality of the mobile power supply meets the standard. After obtaining the texture extension defect depth data, the texture defect geometric distortion analysis is performed. This step aims to conduct a comprehensive distortion assessment of the geometric morphology of the texture disconnection area. First, for the disconnection points at each defect depth, the three-dimensional contour reconstruction technology is used to analyze the geometric shape of the defect area and identify the morphological changes in the fracture area. Subsequently, the geometric distortion detection method is used to analyze the deformation degree and distortion direction of the fracture by comparing the differences between the three-dimensional contours of each defect point and the normal texture structure. This process can identify the geometric deformation caused by extension or stretching, especially those severe distortions in local areas. Finally, the texture defect geometric distortion data is generated to describe the geometric characteristics of the surface defects. After obtaining the texture defect geometric distortion data, the roughness perception assessment of the texture defect is performed. This step uses optical roughness measurement technology, combined with geometric distortion data, to perform a multi-dimensional assessment of the roughness of surface defects. In specific operations, by analyzing the surface reflection characteristics of the distorted area and evaluating the scattering of light on different surface roughnesses, the roughness perception of the texture defect can be determined. Through fine spectral reflection contrast and surface microstructure analysis, the roughness of the surface defect area is identified, and texture defect roughness perception data is generated. This data can reflect the tactile differences of the surface texture defects of the mobile power supply, providing an important reference for the final appearance quality assessment.

[0119] Preferably, step S33 includes the following steps:

[0120] Step S331: performing distribution dispersion analysis on the texture extension defect depth data to obtain defect distribution dispersion data;

[0121] Step S332: Calculate the fault curvature according to the defect distribution dispersion data and the texture extension defect depth data to obtain texture fault curvature data;

[0122] Step S333: performing geometric cluster analysis on the texture fault curvature data to obtain texture fault geometric data;

[0123] Step S334: performing geometric distortion analysis on the texture extension and disconnection data according to the texture fault geometric morphology data to obtain texture defect geometric distortion data.

[0124] In an embodiment of the present invention, the texture extension defect depth data is analyzed for distribution discreteness to identify the spatial distribution characteristics of defects on the surface of the mobile power supply. The operation first uses a statistical method to calculate the discreteness of the spatial position distribution of the defect depth data, and by comparing the distance and depth changes between each defect point, the distribution law of these defects on the surface is obtained. In order to ensure accuracy, the surface of the mobile power supply is divided into several areas by using the surface meshing technology, and the defect depth data in each grid is statistically analyzed to analyze its distribution uniformity and discreteness in the local area. Through this process, it is possible to identify which areas have more concentrated defect distribution and which areas show a larger discreteness, generate defect distribution discreteness data, and provide data support for subsequent fault curvature calculation. According to the defect distribution discreteness data and the texture extension defect depth data, the fault curvature calculation is performed in step S332. First, based on the depth information of each defect point, the surface fitting algorithm is used to reconstruct the geometric morphology of the mobile power supply surface to obtain the preliminary morphology of the surface fault. Subsequently, by calculating the curvature change of the surface at each defect position, the geometric morphology characteristics of the defect area are analyzed, especially the curvature fluctuation of the fracture and depression areas. The continuous change of the fault is analyzed by the difference method, the local extreme point and change trend of the curvature are calculated, and the curvature fluctuation characteristics of the surface geometry are identified. Finally, the texture fault curvature data is obtained, which can accurately describe the curvature change of the surface defects of the mobile power supply and lay the foundation for the subsequent geometric clustering analysis. The geometric clustering analysis of the texture fault curvature data is carried out to identify the geometric feature categories of the surface defects of the mobile power supply. First, the feature vectors of different defect points are extracted based on the curvature data, and the geometric similarity between each defect point is analyzed by the feature distance measurement technology. Subsequently, the hierarchical clustering method is applied to classify the defect points with similar curvature characteristics to form multiple cluster groups. The defect points in each cluster group have similar curvature and depth characteristics. Through this process, the defects on the surface of the mobile power supply can be divided into different types, and the defect areas with common geometric features can be identified. Finally, the texture fault geometric data is obtained, which can clearly characterize the geometric characteristics of the surface defects and facilitate the subsequent geometric distortion analysis. According to the texture fault geometry data, the texture extension disconnection data is subjected to geometric distortion analysis. In this step, the geometric clustering results are used to analyze the degree of distortion of different types of geometric defect areas. First, the defect points in each cluster group are reconstructed in three dimensions to accurately describe their geometric contours. Then, based on the geometric changes in each disconnection area, the surface analysis method is used to calculate the degree of distortion of the surface at the disconnection. This process identifies the direction and degree of distortion of the local area, especially the geometric deformation phenomena such as local bending and stretching, by comparing the changes in surface tension and curvature of the disconnection area.Through this operation, texture defect geometric distortion data is generated to describe the morphological anomalies of the geometric structure of the surface defects of the mobile power supply, providing a basis for subsequent appearance quality evaluation.

[0125] Preferably, performing geometric cluster analysis on texture fault curvature data comprises the following steps:

[0126] Calculating the boundary difference of the texture fault curvature data to obtain the texture fault curvature boundary difference data;

[0127] According to the texture fault curvature boundary difference data, the texture fault curvature data is analyzed for morphological trend, and the texture fault morphological trend data is obtained;

[0128] According to the texture fault curvature boundary difference data and texture fault morphology trend data, geometric morphology cluster analysis is performed to obtain texture fault geometric morphology data.

[0129] In an embodiment of the present invention, the boundary difference degree of the texture fault curvature data is calculated to analyze the curvature boundary characteristics of the fault area. First, the fault curvature change on the surface of the mobile power supply is extracted by using a boundary detection algorithm to generate boundary lines containing different fault areas. Then, the local difference degree of the curvature value on each boundary is calculated, the curvature change amplitude between adjacent points is analyzed, and the boundary difference degree value is obtained by differential calculation. Each boundary point in this process is accurately processed by the curvature comparison method to ensure that subtle curvature changes are also captured. By performing a comprehensive boundary difference degree analysis on the entire surface, texture fault curvature boundary difference data is generated. This data can reflect the curvature difference characteristics of each fault area and provide a key basis for subsequent morphological trend analysis. According to the texture fault curvature boundary difference data, the texture fault curvature data is analyzed for morphological trend. First, the curvature change of each fault area is tracked by curve fitting technology to identify the trend of the curvature change. In this analysis process, a derivative-based curve guidance algorithm is used to deduce the morphological trend of each fault area by calculating the directional gradient of the curvature change. At the same time, the morphological operation method is used to smooth the direction of the curvature boundary to eliminate the error caused by local abnormal changes and ensure the accuracy of the morphological trend analysis. The texture fault morphological trend data finally generated can accurately describe the morphological change direction of each fault area and provide morphological feature support for geometric morphological clustering analysis. Based on the texture fault curvature boundary difference data and texture fault morphological trend data, the fault curvature data is subjected to geometric morphological clustering analysis. In this step, firstly, according to the boundary difference and morphological trend data, the characteristic vectors of each fault area are extracted. These vectors include the curvature change amplitude, trend gradient and local morphological characteristics. Then, the density clustering algorithm is used to cluster the fault areas with similar characteristics, and the fault areas with similar morphological characteristics are assigned to the same cluster group. In order to ensure the accuracy of the clustering results, the Euclidean distance metric is used to further refine the geometric morphological differences between different regions. Finally, the clustering results can accurately divide the texture faults on the surface of the mobile power supply into multiple morphologically similar areas and generate texture fault geometric morphological data. Through this data, the geometric characteristics of surface defects of mobile power supplies can be comprehensively classified and analyzed, providing an accurate basis for subsequent quality inspections.

[0130] Preferably, step S34 includes the following steps:

[0131] Step S341: performing a concave-convex distribution structure analysis on the texture defect geometric distortion data based on the texture extension defect depth data to obtain the texture defect concave-convex distribution data;

[0132] Step S342: performing contact force simulation on the texture defect concave-convex distribution data to obtain the concave-convex distribution contact force data;

[0133] Step S343: calculating the holding friction coefficient for the texture defect concave-convex distribution data according to the concave-convex distribution contact force data to obtain the texture defect holding friction coefficient;

[0134] Step S344: performing roughness perception evaluation on the texture defect geometric distortion data according to the texture defect grip friction coefficient and the concave-convex distribution contact force data to obtain texture defect roughness perception data.

[0135] In an embodiment of the present invention, the texture defect geometric distortion data is analyzed for the concave-convex distribution structure based on the texture extension defect depth data. First, the texture defects on the surface of the mobile power supply are scanned in detail using a high-precision optical measuring instrument to obtain surface height data. These data include the specific location of pits and protrusions and their depth information. Then, the surface height data is processed using morphological filtering technology to remove the influence of noise and perform a structured analysis on the concave-convex surface. Through three-dimensional reconstruction technology, the concave-convex structure is mapped to a two-dimensional plane to form a clear concave-convex distribution map. The figure shows the specific concave and protruding distribution of each defect area, thereby generating texture defect concave-convex distribution data, which provides basic information for subsequent contact force simulation. Contact force simulation is performed on the texture defect concave-convex distribution data. First, a virtual model containing a concave-convex structure is established to simulate the contact situation encountered in actual use. In this process, contact forces of different directions and intensities are applied to simulate the touch and pressure in actual use. The stress distribution of each concave-convex point under different force conditions is calculated using the finite element analysis method. Through this method, the force response data of the concave-convex structure in practical applications, that is, the concave-convex distribution contact force data, can be obtained. These data show the deformation and stress concentration of the concave-convex points during actual use, providing a basis for further calculation of the friction coefficient. The gripping friction coefficient of the concave-convex distribution data of the texture defect is calculated based on the concave-convex distribution contact force data. First, the contact stress and contact area information of each concave-convex area are extracted from the concave-convex distribution contact force data. Then, the friction coefficient of different concave-convex areas is calculated by the friction mechanics formula, which takes into account the relative motion between the contact surfaces and the surface roughness. Using the standard friction coefficient calculation method, the actual force data is compared with the friction mechanics model to obtain the gripping friction coefficient of the texture defect. This coefficient reflects the friction characteristics of the surface texture of the mobile power supply in actual use, thereby providing quantitative data for the rough perception evaluation. The rough perception evaluation of the texture defect geometric distortion data is performed based on the texture defect gripping friction coefficient and the concave-convex distribution contact force data. First, the texture defect gripping friction coefficient is combined with the concave-convex distribution contact force data to comprehensively analyze the surface roughness of each defect area. Using the principles of contact mechanics, the grip friction coefficient is applied to the force data of different concave and convex areas to calculate the roughness perception value of the surface under different conditions. Through calculation, the texture defect roughness perception data is obtained, which shows the perceived roughness of the surface texture in actual use. This evaluation can accurately reflect the actual tactile characteristics of the mobile power supply surface texture and provide detailed roughness information for appearance quality inspection.

[0136] Preferably, step S4 comprises the following steps:

[0137] Step S41: constructing a defect anomaly map for the mobile power supply texture structure reflection map according to the texture defect roughness perception data and the texture defect geometric distortion data to obtain a texture structure defect anomaly map;

[0138] Step S42: performing feature dimensionality reduction on the texture structure defect anomaly map to obtain a texture structure defect anomaly dimensionality reduction map;

[0139] Step S43: Based on the generative adversarial network, the texture structure defect anomaly dimensionality reduction map is classified into defect severity levels to obtain a defect severity level classification report.

[0140] In an embodiment of the present invention, a defect anomaly map is constructed for the mobile power supply texture structure reflection map according to the texture defect roughness perception data and the texture defect geometric distortion data. First, the feature information in the texture defect roughness perception data and the texture defect geometric distortion data is extracted, and these data are combined with the texture structure reflection map by using image processing technology. Then, the defect data is superimposed on the texture structure reflection map by using a differential mapping method to form a defect anomaly map. In this process, the roughness and geometric distortion degree of each defect area are analyzed, and the abnormal area is highlighted by setting a threshold, so as to obtain a texture structure defect anomaly map. The map can show the abnormal area and its distribution in the texture structure in detail, and provide basic data for subsequent feature dimensionality reduction and defect severity classification. The texture structure defect anomaly map is subjected to feature dimensionality reduction processing. First, the principal component analysis (PCA) method is used to process the texture structure defect anomaly map to extract the main feature vectors. These feature vectors include information such as the size, shape and distribution pattern of the defect area. Through dimensionality reduction technology, high-dimensional feature data is compressed to a lower dimension, retaining the most representative feature information. In the process of dimensionality reduction, sufficient variability is retained to ensure that the data after feature dimensionality reduction can still effectively reflect the abnormal situation of the defect. Finally, the texture structure defect anomaly dimensionality reduction map is obtained, which shows the main features of the defect in the space after dimensionality reduction, providing a clear data basis for the subsequent defect severity classification. The texture structure defect anomaly dimensionality reduction map is classified into defect severity levels based on the generative adversarial network (GAN). First, the texture structure defect anomaly dimensionality reduction map is input into the generator of the generative adversarial network, and the generator generates a defect feature map that matches the input map. Then, the feature map is input into the discriminator, which classifies the defects in the map and identifies defects of different severity levels. The training process of the generative adversarial network includes adjusting the network parameters through actual defect data to optimize the classification results. The final defect severity classification report shows the severity of different defect areas and provides a basis for the quality inspection of mobile power supplies. This method can accurately classify and evaluate defects in texture structures and provide detailed reports for appearance quality inspection.

[0141] Preferably, step S5 comprises the following steps:

[0142] Step S51: normalizing the defect severity classification report to obtain a defect severity classification normalized report;

[0143] Step S51: Designing an automated quality monitoring firmware based on the defect severity classification normalization report to obtain an automated quality monitoring firmware;

[0144] Step S52: Send the automated quality monitoring firmware to the cloud platform to perform real-time appearance quality detection.

[0145] In an embodiment of the present invention, the defect severity classification report is normalized. First, each defect area and its corresponding severity data in the defect severity classification report are extracted. Then, normalization is performed, and the specific method is to scale the severity data to a range of 0 to 1. This process achieves data standardization by calculating the ratio of the severity of each defect to the maximum severity value. This normalization process can eliminate the absolute value difference between different defect severity data, so that the data of each defect can be compared at the same scale. Finally, a defect severity classification normalization report is generated, which contains standardized severity data of all defect areas, providing a unified data format and benchmark for subsequent quality monitoring firmware design. Automated quality monitoring firmware design is performed based on the defect severity classification normalization report. First, the defect data in the normalization report is used to design the algorithm logic of the monitoring firmware, including functions such as defect detection, severity judgment and alarm mechanism. In the firmware design process, the thresholds of various defects are defined, and corresponding response strategies are set according to different severity levels. Next, these design schemes are converted into code to realize the functions of the monitoring firmware, including data reception, processing, storage and alarm. Finally, the designed firmware is packaged as an automated quality monitoring firmware to ensure that it can automatically receive and process real-time detection data and trigger the preset alarm mechanism when an abnormality is detected. This firmware can monitor the appearance quality of the production process in real time, and promptly discover and handle quality problems. The automated quality monitoring firmware is sent to the cloud platform to perform real-time appearance quality detection. First, prepare the operation of uploading the firmware to the cloud platform, including setting up the upload interface and ensuring the security of data transmission. Next, upload the automated quality monitoring firmware to the cloud platform through a secure network connection. During the upload process, verify the integrity and correctness of the firmware to ensure that the data has not been tampered with or damaged. After the upload is completed, configure the cloud platform's real-time monitoring system so that the firmware can run on the platform and perform appearance quality detection. The cloud platform will receive real-time data from the production line, use the firmware for quality detection, and generate reports or trigger alarms when defects are detected. This process ensures that the appearance quality of mobile power supplies during production can be monitored and managed in a timely manner.

[0146] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0147] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for detecting the appearance quality of a mobile power source based on machine vision, characterized in that: The following steps are involved: Step S1: using a high-resolution spectral camera to collect multi-angle spectral images of the mobile power supply to obtain a spectral sharpened image of the mobile power supply; Step S2: performing texture structure reflection spectrum analysis on the mobile power supply spectral sharpening image to obtain the mobile power supply texture structure reflection spectrum; performing texture ductility analysis on the mobile power supply texture structure reflection spectrum to obtain texture connection ductility data; wherein step S2 includes: Step S21: using a preset mobile power spectrum standard manual to perform spectrum deviation analysis on the mobile power spectrum sharpening image to obtain mobile power spectrum deviation data; Step S22: performing texture structure reflection spectrum analysis on the mobile power supply spectrum sharpening image according to the mobile power supply spectrum deviation data to obtain a mobile power supply texture structure reflection spectrum; Step S23: identifying connection interaction points of the mobile power supply texture structure reflection map to obtain texture connection interaction point data; Step S24: performing texture ductility analysis on the mobile power supply texture structure reflection map according to the texture connection interaction point data to obtain texture connection ductility data; wherein step S24 includes: Step S241: Calculating the texture interlacing density of different connection interaction points on the mobile power supply texture structure reflection map according to the texture connection interaction point data to obtain the connection interaction point texture interlacing density data; Step S242: performing texture morphology expansion analysis of different connection interaction points on the mobile power supply texture structure reflection map according to the connection interaction point texture interlacing density data to obtain texture morphology expansion data; Step S243: performing high-frequency deformation interference effect analysis on the texture morphology expansion data to obtain high-frequency deformation interference effect data; Step S244: performing texture direction stretching analysis on the texture morphology expansion data according to the high-frequency deformation interference effect data to obtain texture direction stretching data; Step S245: performing texture ductility analysis on the reflection spectrum of the mobile power supply texture structure according to the texture direction stretching data and the high-frequency deformation interference effect data to obtain texture connection ductility data; Step S3: according to the texture connection ductility data, the texture extension and disconnection recognition is performed on the reflection map of the mobile power supply texture structure to obtain the texture extension and disconnection data; the geometric distortion analysis is performed on the texture extension and disconnection data to obtain the texture defect geometric distortion data; the rough perception evaluation is performed on the texture defect geometric distortion data to obtain the texture defect rough perception data; wherein, step S3 includes: Step S31: performing texture extension and disconnection identification on the mobile power supply texture structure reflection map according to the texture connection ductility data to obtain texture extension and disconnection data; Step S32: performing defect depth identification on the texture extension disconnection data to obtain texture extension defect depth data; Step S33: performing geometric distortion analysis on the texture extension disconnection data according to the texture extension defect depth data to obtain texture defect geometric distortion data; wherein step S33 includes: Step S331: performing distribution dispersion analysis on the texture extension defect depth data to obtain defect distribution dispersion data; Step S332: Calculate the fault curvature according to the defect distribution dispersion data and the texture extension defect depth data to obtain texture fault curvature data; Step S333: performing geometric cluster analysis on the texture fault curvature data to obtain texture fault geometric data; wherein step S333 includes the following steps: Calculating the boundary difference of the texture fault curvature data to obtain the texture fault curvature boundary difference data; According to the texture fault curvature boundary difference data, the texture fault curvature data is analyzed for morphological trend, and the texture fault morphological trend data is obtained; According to the texture fault curvature boundary difference data and texture fault morphology trend data, geometric morphology cluster analysis is performed to obtain texture fault geometric morphology data; Step S334: performing geometric distortion analysis on the texture extension and disconnection data according to the texture fault geometric morphology data to obtain texture defect geometric distortion data; Step S34: performing roughness perception evaluation on the texture defect geometric distortion data to obtain texture defect roughness perception data; wherein step S34 includes: Step S341: performing a concave-convex distribution structure analysis on the texture defect geometric distortion data based on the texture extension defect depth data to obtain the texture defect concave-convex distribution data; Step S342: performing contact force simulation on the texture defect concave-convex distribution data to obtain the concave-convex distribution contact force data; Step S343: calculating the holding friction coefficient for the texture defect concave-convex distribution data according to the concave-convex distribution contact force data to obtain the texture defect holding friction coefficient; Step S344: performing roughness perception evaluation on the texture defect geometric distortion data according to the texture defect grip friction coefficient and the concave-convex distribution contact force data to obtain texture defect roughness perception data; Step S4: classifying the defect severity level of the mobile power supply texture structure reflection map according to the texture defect roughness perception data and the texture defect geometric distortion data to obtain a defect severity level classification report; wherein step S4 includes: Step S41: constructing a defect anomaly map for the mobile power supply texture structure reflection map according to the texture defect roughness perception data and the texture defect geometric distortion data to obtain a texture structure defect anomaly map; Step S42: performing feature dimensionality reduction on the texture structure defect anomaly map to obtain a texture structure defect anomaly dimensionality reduction map; Step S43: classifying the defect severity level of the texture structure defect anomaly dimensionality reduction map based on the generative adversarial network to obtain a defect severity level classification report; Step S5: Design an automated quality monitoring firmware based on the defect severity classification report to obtain the automated quality monitoring firmware, and send the automated quality monitoring firmware to the cloud platform to perform real-time appearance quality detection.

2. The method for detecting the appearance quality of a mobile power source based on machine vision according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: using a high-resolution spectral camera to collect multi-angle spectral images of the mobile power supply to obtain a multi-angle spectral image of the mobile power supply; Step S12: performing image contrast optimization processing on the multi-angle spectrum image of the mobile power source to obtain a spectrum optimized image of the mobile power source; Step S13: performing image sharpening processing on the mobile power source spectrum optimization image to obtain a mobile power source spectrum sharpened image.

3. The method for detecting the appearance quality of a mobile power source based on machine vision according to claim 1, characterized in that: Step S5 includes the following steps: Step S51: normalizing the defect severity classification report to obtain a defect severity classification normalized report; Step S51: Designing an automated quality monitoring firmware based on the defect severity classification normalization report to obtain an automated quality monitoring firmware; Step S52: Send the automated quality monitoring firmware to the cloud platform to perform real-time appearance quality detection.

Images