Waste lithium battery classification processing method and system based on image segmentation identification

Abstract

The invention discloses a waste lithium battery classification processing method and system based on image segmentation identification, and relates to the technical field of waste lithium battery recycling, and the method comprises the steps: configuring a high-resolution camera, a lighting device and an ambient light sensor, collecting multi-view images of waste lithium batteries in real time, obtaining a two-dimensional image data set, and carrying out the recognition of the two-dimensional image data set; inputting the two-dimensional image data set into an SIFT algorithm to detect and match feature points, generating a three-dimensional model through an MVS technology based on the matched feature points, and obtaining geometric information of the surface of the battery by using a triangulation method; according to the method, the high-quality three-dimensional model is generated by combining high-resolution multi-view image acquisition with SIFT feature point matching and the MVS technology, and the quality of the model is optimized through denoising, curvature enhancement and dynamic parameter adjustment, so that accurate modeling of a complex geometrical shape is ensured. And the optimized three-dimensional model is analyzed by using the deep convolutional neural network, so that the robustness and accuracy of defect detection are greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of recycling of waste lithium batteries, and in particular to a method and system for classifying and processing waste lithium batteries based on image segmentation and recognition. Background Art

[0002] With the development of technology, the demand for energy recycling has gradually increased, and the recycling of waste lithium batteries has become a hot research field. Traditional recycling relies on manual inspection, which is not only inefficient but also has problems with insufficient inspection accuracy and is prone to misjudgment. In recent years, more and more research has pointed to improving the classification accuracy of waste lithium batteries. By combining multi-view images and three-dimensional reconstruction technology, detailed battery surface information has been provided without damaging the battery surface structure, laying a foundation for defect analysis and subsequent classification and processing.

[0003] However, the existing methods for classifying and processing waste lithium batteries still have some deficiencies. On the one hand, the ability to accurately model complex geometric shapes is limited, resulting in low-quality three-dimensional models. On the other hand, the existing defect detection algorithms lack sufficient robustness and accuracy when dealing with different types of defects, which may lead to key defects being missed or misjudged, affecting the final classification results. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a method for classifying and processing waste lithium batteries based on image segmentation and recognition to solve the problems of poor three-dimensional model quality and insufficient defect detection accuracy.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: In a first aspect, the present invention provides a method for classifying and processing waste lithium batteries based on image segmentation and recognition, which includes configuring a high-resolution camera, a lighting device, and an ambient light sensor to collect multi-view images of waste lithium batteries in real time and obtain a two-dimensional image dataset; Input the two-dimensional image dataset into the SIFT algorithm to detect and match feature points. Based on the matched feature points, generate a three-dimensional model through the MVS technology, and use the triangulation method to obtain the geometric information of the battery surface; Perform denoising and curvature enhancement operations on the three-dimensional model, and dynamically adjust the parameter settings to obtain an optimized three-dimensional model; Use a deep convolutional neural network to analyze the optimized three-dimensional model, identify and locate the defects on the battery, and obtain the position, type, and severity score of the defects; Integrate the location, type, and severity score of the defect and the geometric information of the battery surface to evaluate the battery and obtain the category information and real-time status of the battery; Integrate the optimized three-dimensional model parameters, as well as the category information and real-time status of the battery, generate a classification report, and process the batteries by category based on the classification report.

[0007] As a preferred solution of the waste lithium battery classification and treatment method based on image segmentation recognition according to the present invention, wherein: input the two-dimensional image dataset into the SIFT algorithm to detect and match feature points, and based on the matched feature points, generate a three-dimensional model through the MVS technology, specifically including the following steps, Input the two-dimensional image dataset into the SIFT algorithm, extract multiple scale-invariant feature points, screen out inlier pairs through a FLANN matcher, establish cross-view geometric correspondences, and obtain the matched feature points; Based on the matched feature points, use COLMAP in MVS to calculate the depth information of each pixel, generate a sparse depth map, optimize the point cloud density through point cloud filtering, and finally construct a three-dimensional model through the Marching Cubes algorithm.

[0008] As a preferred solution of the waste lithium battery classification and treatment method based on image segmentation recognition according to the present invention, wherein: the geometric information of the battery surface is obtained by reconstructing a three-dimensional point cloud through multi-view images based on the triangulation method, filtering and density optimizing the three-dimensional point cloud, and then performing operations such as normal vector and curvature calculation and surface deformation feature analysis.

[0009] As a preferred solution of the waste lithium battery classification and treatment method based on image segmentation recognition according to the present invention, wherein: dynamically adjust the parameter settings to obtain an optimized three-dimensional model, specifically including the following steps, Set initial parameters based on the initial point cloud density and curvature distribution of the three-dimensional model, and display the effect of the three-dimensional model under the initial parameters in real time through the voxel grid algorithm; Based on the effect of the three-dimensional model, dynamically fine-tune the initial parameters by reducing the denoising intensity and increasing the curvature sensitivity, further correct the parameters in combination with the effect of the three-dimensional model after parameter adjustment, and at the same time adopt a regional differentiation strategy to automatically assign different parameter weights to high-curvature regions and low-curvature regions; Repeat the process of parameter adjustment until the three-dimensional model meets the accuracy index, lock the final parameters and save the optimized three-dimensional model; The accuracy index is defined based on the geometric difference between the three-dimensional model and the real battery surface.

[0010] As a preferred embodiment of the method for classifying and processing waste lithium batteries based on image segmentation and recognition of the present invention, wherein: the optimized three-dimensional model is analyzed using a deep convolutional neural network to identify and locate defects on the battery, and obtain the position, type, and severity score of the defects. The specific steps are as follows: The optimized three-dimensional model is converted into two-dimensional projection data through multi-view projection, and input into the deep convolutional neural network to extract local curvature features and global shape features, and combine RPN to locate the defect area; Based on the defect area, through the multi-task learning branch of the Faster R-CNN framework, perform the fusion of local curvature features and global shape features and the joint optimization of classification and regression tasks to obtain the bounding box coordinates, type, and severity score of the defect.

[0011] As a preferred embodiment of the method for classifying and processing waste lithium batteries based on image segmentation and recognition of the present invention, wherein: the position, type, and severity score of the defect and the geometric information on the battery surface are integrated to evaluate the battery and obtain the category information and real-time status of the battery. The specific steps are as follows: The position, type, and severity score of the defect and the geometric information on the battery surface are mapped back to the three-dimensional space through coordinate transformation, associated with the same data structure using point cloud spherical hashing mapping, and combined with the KD-Tree spatial index to generate the integrated information; Based on the integrated information, evaluate the surface defect pattern of the battery through a convolutional neural network to obtain the category information of the battery, and evaluate the dynamic charge and discharge behavior of the battery through a lightweight temporal autoencoder to obtain the real-time status of the battery.

[0012] As a preferred embodiment of the method for classifying and processing waste lithium batteries based on image segmentation and recognition of the present invention, wherein: the classification report is generated through multi-dimensional data aggregation and document automation tools based on the integrated three-dimensional model parameters, category information, and real-time status. Based on the risk level label in the classification report, use an automated decision tree to trigger the classification, processing, and disposal process of waste lithium batteries.

[0013] In a second aspect, the present invention provides a system for classifying and processing waste lithium batteries based on image segmentation and recognition, including an image acquisition module, a three-dimensional reconstruction module, a three-dimensional model optimization module, a defect detection and location module, a comprehensive evaluation module, and a classification processing module; The image acquisition module is used to configure a high-resolution camera, lighting equipment, and an ambient light sensor to collect multi-view images of waste lithium batteries in real time and obtain a two-dimensional image dataset; A 3D reconstruction module for inputting a 2D image dataset into the SIFT algorithm to detect and match feature points, generating a 3D model based on the matched feature points by means of the MVS technology, and obtaining the geometric information of the battery surface using the triangulation method; A 3D model optimization module for performing denoising and curvature enhancement operations on the 3D model and dynamically adjusting parameter settings to obtain an optimized 3D model; A defect detection and localization module for analyzing the optimized 3D model using a deep convolutional neural network to identify and locate defects on the battery, and obtaining the position, type, and severity score of the defects; A comprehensive evaluation module for integrating the position, type, and severity score of the defects and the geometric information of the battery surface to evaluate the battery and obtain the category information and real-time status of the battery; A classification processing module for integrating the parameters of the optimized 3D model, as well as the category information and real-time status of the battery, generating a classification report, and processing the batteries by category based on the classification report.

[0014] In a third aspect, the present invention provides a computer device, including a memory and a processor, where the memory stores a computer program, and: when the computer program is executed by the processor, any step of the waste lithium battery classification processing method based on image segmentation recognition as described in the first aspect of the present invention is implemented.

[0015] In a fourth aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored, and: when the computer program is executed by the processor, any step of the waste lithium battery classification processing method based on image segmentation recognition as described in the first aspect of the present invention is implemented.

[0016] The beneficial effects of the present invention are as follows: High-resolution multi-view image acquisition is combined with SIFT feature point matching and MVS technology to generate a high-quality 3D model, and the model quality is optimized through denoising, curvature enhancement, and dynamic parameter adjustment, ensuring accurate modeling of complex geometric shapes. Secondly, a deep convolutional neural network is used to analyze the optimized 3D model, and the RPN and Faster R-CNN frameworks are combined to achieve the fusion of local curvature features and global shape features and the joint optimization of classification and regression tasks, greatly improving the robustness and accuracy of defect detection. It can not only accurately locate and classify various types of defects, but also accurately evaluate their severity, thus providing strong support for the precise classification and efficient recycling of waste lithium batteries. Description of the Drawings

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0018] Figure 1 It is a flowchart of the method for classifying and processing waste lithium batteries based on image segmentation and recognition in Embodiment 1.

[0019] Figure 2 It is a schematic diagram of the system for classifying and processing waste lithium batteries based on image segmentation and recognition in Embodiment 1.

[0020] Figure 3 It is a flowchart of the method for classifying and processing waste lithium batteries in Embodiment 1.

[0021] Figure 4 It is a flowchart of the structure of the system for classifying and processing waste lithium batteries in Embodiment 1. Specific Embodiments

[0022] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following will make a detailed description of the specific embodiments of the present invention with reference to the accompanying drawings of the specification.

[0023] In the following description, many specific details are set forth to fully understand the present invention. However, the present invention can also be implemented in other ways different from those described herein. Those skilled in the art can make similar extensions without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0024] Secondly, the so-called "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that can be included in at least one implementation manner of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment that excludes other embodiments.

[0025] Embodiment 1, referring to Figures 1 to 4 , is the first embodiment of the present invention. This embodiment provides a method for classifying and processing waste lithium batteries based on image segmentation and recognition, including the following steps: S1. Configure a high-resolution camera, lighting equipment, and an ambient light sensor to collect multi-view images of waste lithium batteries in real time and obtain a two-dimensional image dataset.

[0026] Specifically, the operations are as follows. To achieve multi - perspective image acquisition of waste lithium batteries, detailed hardware configuration and calibration work need to be carried out first. High - resolution industrial cameras are placed at different positions and angles to ensure that all surfaces of the battery can be covered. Each camera is equipped with independently controlled LED lighting equipment, which ensures clear and shadow - free images even under changing ambient light conditions. At the same time, an ambient light sensor monitors the ambient light conditions in real - time and automatically adjusts the brightness and color temperature of the lighting equipment according to the ambient light information, thereby maintaining the consistency of imaging quality; Next, precise calibration and synchronous data acquisition of the cameras are carried out. Each camera is calibrated using a standard checkerboard pattern to determine its focal length, principal point, and the position and orientation of the camera relative to the battery. This step is crucial because it directly affects the accuracy of subsequent 3D reconstruction. All cameras need to work synchronously. The global shutter technology is used to capture images to avoid blurring caused by motion. Synchronization can be achieved through GPIO signals and the Network Time Protocol (NTP) to ensure that all cameras capture images at the same moment and guarantee the consistency of image data. Each time a shot is taken, preliminary pre - processing such as denoising and contrast enhancement is performed through the in - camera ISP to improve the image quality; After completing multi - perspective image acquisition, the acquired two - dimensional image data is transmitted to the central processing unit in real - time through a high - speed network interface. To ensure the integrity and security of the two - dimensional image data, redundant array of independent disks (RAID) technology is used for storing the two - dimensional image data to prevent data loss caused by hardware failures. In addition, HDFS can be used to efficiently manage large - scale two - dimensional image data, and a dedicated data management framework is developed to support indexing, retrieval, and version control of image data, facilitating technicians to view and manage the acquired two - dimensional image dataset.

[0027] S2: Input the two - dimensional image dataset into the SIFT algorithm to detect and match feature points. Based on the matched feature points, generate a 3D model through the MVS technology and obtain the geometric information of the battery surface using the triangulation method.

[0028] Specifically, the following operations are included, In the initial stage of waste lithium battery classification and treatment, the multi - perspective two - dimensional images collected need to be processed. The core of this process is to use the SIFT (Scale - Invariant Feature Transform) algorithm to construct Gaussian pyramids at different scales and find local extreme points as feature points through differential calculation. First, the original two - dimensional image is continuously scaled and a Gaussian filter is applied to generate a series of smoothed images, forming a Gaussian pyramid. The difference calculation between adjacent layers forms a difference pyramid, which is used to detect local maximum and minimum points as candidate feature points. This process can not only capture the tiny details on the battery surface but also effectively avoid mis - matching problems caused by light changes or texture differences; To ensure the stability and uniqueness of feature points, it is crucial to further remove low-contrast points and perform edge response filtering. The exact position and response value of each candidate feature point are approximately calculated using Taylor expansion, and low-contrast points are eliminated. At the same time, the principal curvature ratio is calculated using the Hessian matrix to remove unstable points located on the edges. Next, one or more principal directions are assigned to each feature point, and the peak value is selected as the principal direction by calculating the gradient direction histogram to ensure rotational invariance. This step enables the feature points to remain consistent even when the image is rotated. Finally, the FLANN (Fast Library for Approximate Nearest Neighbors) matcher is used to match these feature points, and inlier pairs are screened out to establish cross-view geometric correspondences, thereby obtaining the matched feature points. The FLANN algorithm significantly improves the matching efficiency and accuracy through fast library approximate nearest neighbor search; Based on the obtained matched feature points, the COLMAP software in the MVS (Multi-View Stereo) technology is used to calculate the depth information of each pixel and generate a sparse depth map. COLMAP jointly optimizes the camera pose and the position of 3D points through bundle adjustment technology to further improve the reconstruction accuracy. This process greatly enhances the accuracy of 3D reconstruction, especially in capturing details of complex shapes such as the battery surface. The bundle adjustment technology globally optimizes all camera parameters and the positions of 3D points to ensure geometric consistency; After generating the sparse depth map, it is necessary to filter and optimize its density. Commonly used filtering methods include statistical filtering and voxel gridding, which are used to remove noise points and increase the point cloud density. Statistical filtering removes isolated noise points by analyzing the number of neighbor points around each point, while voxel gridding divides the point cloud into uniform small cube cells and interpolates the points within each cell to supplement missing points, thereby improving the overall density of the point cloud. Subsequently, the Marching Cubes algorithm is used to convert the optimized point cloud into a triangular mesh. By traversing the voxel space to identify cube cells containing the surface and generating the corresponding triangular mesh, a complete 3D model is finally constructed. The Marching Cubes algorithm can not only accurately represent complex surface morphologies but also generate a smooth and coherent 3D model, making the 3D model both aesthetically pleasing and practical; In addition, Laplacian smoothing is applied to reduce the number of polygons while maintaining the visual effect. Laplacian smoothing reduces sharp edges in the 3D model by weighted averaging the positions of each vertex, enhancing the visual fineness. This is crucial for subsequent analysis and applications. The entire 3D model generation process not only improves the reliability and accuracy of the processing results but also provides strong technical support for subsequent defect detection, classification tasks, and the recycling and reuse of waste lithium batteries; After the 3D model is constructed, the geometric information on the battery surface is further extracted. First, according to the triangulation method, a 3D point cloud is reconstructed from multi-view images, and the ICP (Iterative Closest Point) algorithm is used to register the point clouds from different views to ensure seamless docking between the point cloud parts. The ICP algorithm gradually optimizes the alignment accuracy of the point cloud by minimizing the distance error between the point clouds. At the same time, the 3D point cloud is filtered to remove noise, and techniques such as voxel gridding and Poisson reconstruction are used to optimize the point cloud density and smoothness, making the finally generated 3D model more accurate and consistent; Next, the normal vector and curvature of each point in the 3D point cloud are calculated. These properties are crucial for understanding the shape and structure of the battery surface. The normal vector provides the direction information of the surface, while the curvature reflects the degree of surface curvature. Combining the normal vector and curvature information, surface deformation feature analysis is carried out to identify possible uneven or deformed areas. For example, based on the significance of the curvature change in the surface geometry characteristics, a curvature mutation threshold is defined. By analyzing the joint characteristics of the curvature mutation threshold and the surface reflection intensity, defect types such as corrosion and cracks can be accurately located. This in-depth surface deformation feature analysis combined with a convolutional neural network can efficiently classify and identify surface anomalies, providing a scientific basis for the safety assessment and recycling of used lithium batteries; Through the accurate understanding of the surface morphology, the efficiency and reliability of automated processing can be significantly improved. For example, in the operation of a robotic arm, an accurate 3D model and surface information can help the robotic arm grasp and process used batteries more accurately, avoiding damage and omission. This efficient automated processing not only improves work efficiency but also reduces the need for manual intervention and lowers costs.

[0029] S3. Perform denoising and curvature enhancement operations on the 3D model, and dynamically adjust the parameter settings to obtain an optimized 3D model.

[0030] Specifically, the following operations are included. Based on the initial point cloud density and curvature distribution of the 3D model, initial parameters are set. The initial parameters include denoising intensity, curvature sensitivity, etc., which directly affect the quality of the 3D model. In this stage, the voxel grid algorithm is used to convert the original point cloud into a voxel representation. Voxel gridding is a method of dividing 3D space into uniform voxels, calculating the average value or interpolation in each voxel, and thus generating a visual grid model. This method can not only quickly display the general shape of the 3D model but also help evaluate the effectiveness of the initial parameters. For example, for high-density point cloud areas, a smaller voxel size can be used to capture details, while for low-density areas, a larger voxel size can be used to reduce computational complexity. In this way, a relatively clear view of the 3D model can be obtained in the preliminary stage, and the effect of the 3D model under the initial parameters can be displayed in real time, providing an intuitive basis for further parameter adjustment; Dynamically fine-tune the parameters based on the real-time display effect. In specific operations, observe the changes of the 3D model by gradually reducing the denoising intensity and increasing the curvature sensitivity, and further correct the parameters in combination with the model effect after parameter adjustment. This iterative adjustment process aims to find the optimal parameter combination so that the model can not only effectively remove noise but also accurately capture the detailed features of the battery surface. In this process, it is particularly important to use the regional differentiation strategy. Since there are various different geometric characteristics on the surface of waste lithium batteries, such as edges, sharp parts, and flat areas, it is necessary to assign different parameter weights according to the characteristics of different regions. For example, in high-curvature regions, such as the root of the battery tab or indentations, increase the curvature sensitivity to better capture subtle changes; while in low-curvature regions, such as large flat surfaces, appropriately reduce the denoising intensity to retain more detailed information. This can ensure that the 3D model can achieve the best performance in different regions and improve the overall quality. In addition, an adaptive algorithm can be used to automatically identify high-curvature and low-curvature regions and dynamically adjust the parameters according to actual needs to further improve the robustness and adaptability of the 3D model; After multiple parameter adjustments, the 3D model gradually approaches the optimum. Next, it is necessary to verify whether it meets the predefined accuracy indicators. These accuracy indicators are defined based on the geometric differences between the 3D model and the real battery surface and usually include quantitative indicators such as surface error and shape deviation. To accurately evaluate the accuracy of the 3D model, it can be verified through various methods. For example, use a laser scanner or other high-precision measurement equipment to obtain the real geometric data of the battery surface, and then compare and analyze the geometric data with the 3D model. Common evaluation indicators include root mean square error and maximum absolute error. By calculating and analyzing these indicators, the difference degree between the 3D model and the actual battery surface can be comprehensively understood. If all indicators of the 3D model meet the requirements, lock the current parameter settings and save the optimized 3D model. This process not only ensures the high quality of the 3D model but also provides a solid foundation for subsequent defect detection and analysis; Through the above operations, a complete process from the initial parameter setting to the final optimization of the 3D model is achieved, which not only improves the reliability and accuracy of the processing results but also provides high-quality data support for subsequent deep convolutional neural network analysis, significantly enhancing the efficiency and effect of waste lithium battery recycling and reuse.

[0031] S4. Use a deep convolutional neural network to analyze the optimized 3D model, identify and locate the defects on the battery, and obtain the position, type, and severity score of the defects.

[0032] Specifically, the operations are as follows. Before starting to use the deep convolutional neural network for analysis, the optimized 3D model needs to be converted into 2D projection data through multi-view projection. In the specific operation, multiple viewpoints are selected, such as top view, side view, etc., and the 3D model is rendered from these viewpoints to generate the corresponding 2D images. The 2D image under each viewpoint not only contains the basic information of the battery surface, but also retains the key geometric features of the original 3D model. In order to further enhance the quality of the 2D image, some preprocessing techniques can be applied, such as contrast adjustment, edge enhancement, etc., to ensure that the data input to the deep convolutional neural network has high clarity and detail expression; The two-dimensional projection data is input into the convolutional neural network model for feature extraction. Here, pre-trained convolutional neural network architectures such as ResNet and VGG are usually used. These convolutional neural network models have been fully trained on large-scale image datasets and can effectively capture complex features in images. Through forward propagation, the convolutional neural network model automatically extracts multi-level feature representations in the two-dimensional image, including low-level edges, texture information, and high-level semantic features. Then, the RPN (region proposal network) is used to preliminarily locate potential defective areas. The RPN scans features through a sliding window mechanism to generate a series of candidate regions, and then identifies defective areas. The design of the RPN reduces false positives while maintaining a high recall rate, providing a reliable basis for subsequent accurate battery classification; Through the Faster R-CNN framework, a multi-task learning branch is set in the convolutional neural network model, which mainly includes the fusion of local curvature features and global shape features, as well as the joint optimization of classification and regression tasks. First, for each defect area generated by RPN, the local features inside it are further extracted. Specifically, the defect areas of different sizes can be unified to a fixed size through the RoI pooling layer to facilitate subsequent processing. On this basis, the local curvature features and global shape features are combined to form a richer feature representation. The local curvature features can be calculated from the three-dimensional model, reflecting the curvature of the surface in a specific defect area, while the global shape features describe the general shape of the entire battery surface. The fusion of local curvature features and global shapes helps to improve the recognition ability of complex defects; Next, in the joint optimization stage of the classification and regression tasks, the Faster R-CNN framework performs two tasks simultaneously: one is to classify each defect area to determine whether it belongs to a certain type of defect; the other is to perform regression on the bounding boxes to accurately adjust the position and size of the candidate boxes to better match the actual defect positions. The classification task usually uses the Softmax function to output the probability distribution of each defect category, while the regression task optimizes the parameters by minimizing the distance error between the predicted box and the ground truth box. This multi-task learning method not only improves the detection accuracy but also effectively reduces the complexity and computational burden. Finally, the convolutional neural network model based on the Faster R-CNN framework after training and optimization can output the bounding box coordinates, types, and severity scores of each defect area, providing an accurate basis for subsequent battery evaluation; Through the above operations, it is ensured that the optimized three-dimensional model can accurately identify and locate various defects on the battery and obtain detailed defect information.

[0033] S5. Integrate the position, type, and severity score of the defect with the geometric information on the battery surface to evaluate the battery and obtain the category information and real-time status of the battery.

[0034] Specifically, the following operations are included. Convert the defect positions, types, and severity scores detected in the two-dimensional projection back to the corresponding positions in the three-dimensional space. In the specific operation, using the camera parameters and viewpoint information recorded during the previous multi-view projection, through inverse perspective transformation, map each defect position from the two-dimensional image coordinate system back to the coordinate system of the three-dimensional model. This process ensures that all defect information can be accurately reflected on the original three-dimensional model. Next, use the point cloud spherical hashing mapping technology to associate this defect information with the geometric information on the battery surface. Point cloud spherical hashing mapping is an efficient point cloud indexing method that divides the space into multiple spherical regions and assigns a unique hash value to each spherical region to achieve fast lookup and association, and can efficiently integrate the defect information with the point cloud data on the battery surface into the same data structure; To further improve the data query efficiency and processing speed, the KD-Tree (K-Dimensional Tree) spatial indexing technology is adopted. The KD-Tree is a tree structure used to organize multi-dimensional data and is particularly suitable for nearest neighbor search in high-dimensional spaces. By constructing the KD-Tree, the point cloud data closest to a certain defect location can be quickly retrieved, so as to more accurately associate the defect with the specific location on the battery surface. When generating the integrated defect location information, the data after coordinate transformation and point cloud spherical hashing mapping is input into the KD-Tree to form a unified data structure. This data structure not only contains the locations, types, and severity scores of all defects, but also combines the geometric features of the battery surface, such as curvature, normal vector, etc., providing a comprehensive information basis for subsequent evaluation; Based on the integrated defect location information, the surface defect pattern of the battery is evaluated through a convolutional neural network. In the specific operation, the integrated information is input into the convolutional neural network model. The convolutional neural network model will automatically extract multi-level feature representations and use these features for classification tasks to determine the category corresponding to each defect area, such as corrosion, crack, etc. Finally, the convolutional neural network model outputs the category information of the battery, that is, determines which type of problem the battery belongs to according to the surface defect pattern, such as whether it needs to be repaired or scrapped; In addition to the evaluation of the surface defect pattern, it is also necessary to analyze the dynamic charge and discharge behavior of the battery to obtain the real-time state of the battery. The LSTM of the lightweight time series autoencoder is used for evaluation. The LSTM is a recurrent neural network specifically designed to process sequence data and is very suitable for capturing long-term dependencies in time series. In the specific operation, data such as voltage, current, and temperature of the battery in different charge and discharge cycles are collected to form a time series data set, and then these time series data are input into the LSTM Autoencoder for training. The LSTM Autoencoder consists of an encoder and a decoder. The encoder is responsible for compressing the input time series into a low-dimensional representation, and the decoder tries to reconstruct the original time series from this low-dimensional representation. By minimizing the reconstruction error, the LSTM model can learn effective feature representations of the battery charge and discharge behavior. Finally, the trained LSTM Autoencoder is used to predict new time series data, evaluate the current health state of the battery, and give a health score. Its mathematical expression is: ; where, represents the health score, represents the number of time series segments, represents the index of the time series segment, represents the weight of the represents the reconstruction error of the n-th time series segment, represents the mean of all reconstruction errors, represents the number of battery characteristics, represents the index of the battery characteristic, represents the status score of the n-th battery characteristic; The health score helps to identify whether there are potential functional problems with the battery, such as capacity decay and increased internal resistance; Through the above operations, the defect information is efficiently integrated with the battery surface geometry data, and comprehensive evaluation is carried out using a convolutional neural network and a lightweight temporal autoencoder, significantly improving the accuracy and efficiency of the classification and status evaluation of waste lithium batteries.

[0035] S6. Integrate the optimized three-dimensional model parameters, the category information and the real-time status of the battery, generate a classification report, and process the batteries by category based on the classification report.

[0036] Specifically, the following operations are included. To achieve efficient classification and processing of waste lithium batteries, first, it is necessary to integrate the optimized three-dimensional model parameters, the category information and the real-time status of the battery, and generate a detailed classification report. This process starts with multi-dimensional data aggregation, extracts relevant data from various data sources through ETL tools, including three-dimensional model parameters, the location and type of defects, severity scores, and data on battery surface geometry information. After cleaning and transformation, these data are loaded into a comprehensive database to ensure seamless integration of data from different sources. Next, document automation tools are used to automatically generate classification reports according to predefined templates, which usually contain fields in standard formats for presenting various evaluation results, such as defect location, health score, risk level, etc. This can not only improve the efficiency of report generation but also ensure the consistency and accuracy of the format; In the classification report, each battery will be assigned a risk level label based on its defect type, health score, and real-time status evaluation. To determine this risk level label, a scoring system is designed to assign weights to each factor and calculate the total score to determine the final risk level. For example, the risk level can be determined comprehensively based on factors such as the severity and location of the defect and the health status of the battery. Once the risk level is determined, an automated decision tree can be used to trigger the corresponding processing flow. The automated decision tree is a rule-based decision support strategy that can automatically select the appropriate operation path according to the input conditions. For example, for batteries with a high risk level, the decision tree may recommend direct scrapping, while for batteries with a low risk level, it may recommend further inspection or simple repair; Based on the risk level labels in the classification report, the automated decision tree initiates the corresponding classification treatment and disposal process for used lithium batteries. First, a series of rules are configured in the decision tree strategy. These rules define the treatment measures corresponding to different risk levels. For example, set rules such as "if the risk level is high, trigger the scrapping process; if the risk level is medium, enter the repair process; if the risk level is low, mark it as reusable". When the classification report is generated and the risk level is labeled, the automated decision tree automatically triggers the corresponding treatment process according to the preset rules. This can be achieved through the ERP to ensure that the treatment instructions can be quickly conveyed to relevant departments and equipment. During the battery classification treatment process, monitoring equipment is configured to track the treatment progress of each battery in real time and record the treatment results. In addition, a feedback mechanism can be set up to collect the actual treatment effects to continuously optimize the rules and processes in the decision tree; By performing the above operations to generate a detailed classification report and triggering the corresponding treatment process using the automated decision tree based on the risk level labels, the efficiency and accuracy of the classification treatment of used lithium batteries are significantly improved. It not only reduces the possibility of human errors but also enhances the effectiveness of resource recovery, providing strong support for achieving efficient management and environmental protection treatment.

[0037] This embodiment also provides a classification treatment system for used lithium batteries based on image segmentation recognition, including: an image acquisition module, a 3D reconstruction module, a 3D model optimization module, a defect detection and localization module, a comprehensive evaluation module, and a classification treatment module; The image acquisition module is used to configure a high-resolution camera, lighting equipment, and an ambient light sensor to collect multi-view images of used lithium batteries in real time and obtain a 2D image dataset; The 3D reconstruction module is used to input the 2D image dataset into the SIFT algorithm to detect and match feature points. Based on the matched feature points, a 3D model is generated through the MVS technology, and the geometric information of the battery surface is obtained using the triangulation method; The 3D model optimization module is used to perform denoising and curvature enhancement operations on the 3D model and dynamically adjust the parameter settings to obtain an optimized 3D model; The defect detection and localization module is used to analyze the optimized 3D model using a deep convolutional neural network to identify and locate the defects on the battery, and obtain the position, type, and severity score of the defects; The comprehensive evaluation module is used to integrate the position, type, and severity score of the defects and the geometric information of the battery surface to evaluate the battery and obtain the category information and real-time status of the battery; The classification treatment module integrates the optimized 3D model parameters, as well as the category information and real-time status of the battery, generates a classification report, and processes the batteries by category based on the classification report.

[0038] This embodiment also provides a computer device applicable to the case of the waste lithium battery classification and processing method based on image segmentation recognition, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the waste lithium battery classification and processing method based on image segmentation recognition as proposed in the above embodiment.

[0039] This computer device can be a terminal. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected through a system bus. Among them, the processor of this computer device is used to provide computing and control capabilities. The memory of this computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of this computer device is used to communicate with an external terminal in a wired or wireless manner. The wireless manner can be implemented through WIFI, a carrier network, NFC (Near Field Communication), or other technologies. The display screen of this computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of this computer device can be a touch layer covering the display screen, or a button, a trackball, or a touchpad provided on the housing of the computer device, or an external keyboard, a touchpad, or a mouse, etc.

[0040] This embodiment also provides a storage medium, on which a computer program is stored. When the program is executed by a processor, it implements the waste lithium battery classification and processing method based on image segmentation recognition as proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM for short), Electrically Erasable Programmable Read-Only Memory (EEPROM for short), Erasable Programmable Read Only Memory (EPROM for short), Programmable Red-Only Memory (PROM for short), Read-Only Memory (ROM for short), magnetic memory, flash memory, a magnetic disk, or an optical disc.

[0041] In summary, the present invention: generates a high-quality three-dimensional model by using high-resolution multi-view image acquisition in combination with SIFT feature point matching and MVS technology, and optimizes the model quality through denoising, curvature enhancement, and dynamic parameter adjustment, ensuring accurate modeling of complex geometric shapes. Secondly, the optimized three-dimensional model is analyzed using a deep convolutional neural network, and the fusion of local curvature features and global shape features and the joint optimization of classification and regression tasks are achieved in combination with the RPN and Faster R-CNN frameworks, greatly improving the robustness and accuracy of defect detection. It can not only accurately locate and classify various types of defects, but also accurately evaluate their severity, thus providing strong support for the precise classification and efficient recycling of waste lithium batteries.

[0042] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. A method for classifying and processing waste lithium batteries based on image segmentation recognition, characterized in that: Including, Configuring a high-resolution camera, a lighting device, and an ambient light sensor to collect multi-view images of used lithium batteries in real time and obtain a two-dimensional image dataset; Inputting the two-dimensional image dataset into the SIFT algorithm to detect and match feature points, generating a three-dimensional model based on the matched feature points through the MVS technology, and obtaining the geometric information of the battery surface using the triangulation method; Performing denoising and curvature enhancement operations on the three-dimensional model and dynamically adjusting the parameter settings to obtain an optimized three-dimensional model; Using a deep convolutional neural network to analyze the optimized three-dimensional model, identifying and locating defects on the battery, and obtaining the position, type, and severity score of the defects; Integrating the position, type, and severity score of the defects and the geometric information of the battery surface to evaluate the battery and obtain the category information and real-time status of the battery; Integrating the parameters of the optimized three-dimensional model, as well as the category information and real-time status of the battery, generating a classification report, and processing the batteries by category based on the classification report.

2. The method for classifying and processing waste lithium batteries based on image segmentation recognition according to claim 1, wherein: The step of inputting the two-dimensional image dataset into the SIFT algorithm to detect and match feature points and generating a three-dimensional model based on the matched feature points through the MVS technology specifically includes the following steps: Inputting the two-dimensional image dataset into the SIFT algorithm, extracting multiple scale-invariant feature points, screening out inlier pairs through a FLANN matcher, establishing cross-view geometric correspondences, and obtaining the matched feature points; Based on the matched feature points, using COLMAP in MVS to calculate the depth information of each pixel, generating a sparse depth map, optimizing the point cloud density through point cloud filtering, and finally constructing a three-dimensional model through the Marching Cubes algorithm.

3. The method for classifying and processing waste lithium batteries based on image segmentation recognition according to claim 1, wherein: The geometric information of the battery surface is obtained by reconstructing a three-dimensional point cloud from multi-view images based on the triangulation method, filtering and optimizing the density of the three-dimensional point cloud, and then performing operations such as normal vector and curvature calculation and surface deformation feature analysis.

4. The method for classifying and processing waste lithium batteries based on image segmentation recognition according to claim 3, wherein: The step of dynamically adjusting the parameter settings to obtain an optimized three-dimensional model specifically includes the following steps: Setting initial parameters based on the initial point cloud density and curvature distribution of the three-dimensional model, and displaying the effect of the three-dimensional model under the initial parameters in real time through the voxel grid algorithm; Based on the effect of the three-dimensional model, dynamically fine-tuning the initial parameters by reducing the denoising intensity and increasing the curvature sensitivity, further correcting the parameters in combination with the effect of the three-dimensional model after parameter adjustment, and at the same time adopting a regional differentiation strategy to automatically assign different parameter weights to high-curvature regions and low-curvature regions; Repeating the process of parameter adjustment until the three-dimensional model meets the accuracy index, locking the final parameters and saving the optimized three-dimensional model; The accuracy index is defined based on the geometric difference between the three-dimensional model and the real battery surface.

5. The method for classifying and processing waste lithium batteries based on image segmentation recognition according to claim 4, characterized in that: The step of using a deep convolutional neural network to analyze the optimized three-dimensional model, identifying and locating defects on the battery, and obtaining the position, type, and severity score of the defects specifically includes the following steps: Converting the optimized three-dimensional model into two-dimensional projection data through multi-view projection, inputting it into a deep convolutional neural network to extract local curvature features and global shape features, and combining RPN to locate the defect area; Based on the defective area, through the multi-task learning branch of the Faster R-CNN framework, the fusion of local curvature features and global shape features and the joint optimization of classification and regression tasks are carried out to obtain the bounding box coordinates, type and severity score of the defect.

6. The method for classifying and processing waste lithium batteries based on image segmentation recognition according to claim 5, wherein: Integrate the location, type and severity score of the defect and the geometric information of the battery surface, evaluate the battery, and obtain the category information and real-time status of the battery. The specific steps are as follows. Map the location, type and severity score of the defect and the geometric information of the battery surface back to the three-dimensional space through coordinate transformation, associate them to the same data structure using point cloud spherical hashing mapping, and combine with the KD-Tree spatial index to generate the integrated information. Based on the integrated information, evaluate the surface defect pattern of the battery through a convolutional neural network to obtain the category information of the battery, and evaluate the dynamic charge and discharge behavior of the battery through a lightweight temporal autoencoder to obtain the real-time status of the battery.

7. The method for classifying and processing waste lithium batteries based on image segmentation recognition according to claim 6, characterized in that: The processing of the battery by category based on the classification report includes generating a classification report through multi-dimensional data aggregation and document automation tools based on the integrated three-dimensional model parameters, category information and real-time status, and triggering the classification processing and disposal process of waste lithium batteries using an automated decision tree based on the risk level label in the classification report.

8. A waste lithium battery classification and treatment system based on image segmentation recognition, based on the waste lithium battery classification and treatment method based on image segmentation recognition according to any one of claims 1 to 7, characterized in that: It includes an image acquisition module, a 3D reconstruction module, a 3D model optimization module, a defect detection and localization module, a comprehensive evaluation module, and a classification processing module. The image acquisition module is used to configure a high-resolution camera, lighting equipment and an ambient light sensor to collect multi-view images of waste lithium batteries in real time and obtain a two-dimensional image dataset. The 3D reconstruction module is used to input the two-dimensional image dataset into the SIFT algorithm to detect and match feature points, generate a 3D model based on the matched feature points, and obtain the geometric information of the battery surface using triangulation. The 3D model optimization module is used to perform denoising and curvature enhancement operations on the 3D model and dynamically adjust parameter settings to obtain an optimized 3D model. The defect detection and localization module is used to analyze the optimized 3D model using a deep convolutional neural network to identify and locate the defects on the battery and obtain the location, type and severity score of the defects. The comprehensive evaluation module is used to integrate the location, type and severity score of the defect and the geometric information of the battery surface, evaluate the battery, and obtain the category information and real-time status of the battery. The classification processing module integrates the optimized 3D model parameters and the category information and real-time status of the battery, generates a classification report, and processes the battery by category based on the classification report.

9. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the waste lithium battery classification processing method based on image segmentation recognition according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the waste lithium battery classification processing method based on image segmentation recognition according to any one of claims 1 to 7.

Images