A control method and system for three-electrode battery cell assembly equipment based on image recognition

Abstract

The present invention relates to the technical field of three-pole battery cell assembly, and in particular to a three-pole battery cell assembly equipment control method and system based on image recognition, which obtain an actual assembly state image of a three-pole battery cell at a preset time node, perform feature extraction processing on the actual assembly state image; discretize the contour curve of the three-pole battery cell, perform redundancy correction processing on the discrete points obtained by discretization, and construct an actual assembly state model diagram of the three-pole battery cell according to the corrected discrete points; calculate the overlap degree between the actual assembly state model diagram and the standard assembly state model diagram; analyze the assembly state of the three-pole battery cell according to the overlap degree between the actual assembly state model diagram and the standard assembly state model diagram, and regulate the three-pole battery cell assembly equipment according to the analysis result, which can significantly improve the assembly accuracy, reduce the scrap rate, and realize the intelligent and refined control of the production process.

Description

Technical Field

[0001] The present invention relates to the technical field of three-pole-tab battery cell assembly, and in particular to a three-pole-tab battery cell assembly equipment control method and system based on image recognition. Background Art

[0002] With the rapid development of the electronics industry, three-pole batteries are playing an increasingly important role in electric vehicles, drones, mobile devices and other fields. The traditional three-pole battery assembly control process usually relies on mechanical sensing and hard-coded rules, but it is difficult to adapt to the diversity of batteries under complex working conditions. The development of image recognition technology, especially the advancement of deep learning, enables computers to extract and understand the features of objects from images, providing new solutions for battery assembly. By capturing real-time images of batteries through cameras and using advanced image processing algorithms such as edge detection, contour tracking and feature matching, the real-time assembly status of three-pole batteries and other information can be accurately identified. This information is transmitted to the control system in real time, which can accurately control the assembly process of the assembly equipment and realize intelligent control, thereby greatly improving the accuracy and speed of assembly and promoting the intelligent manufacturing process of the battery manufacturing industry.

[0003] Although the image recognition-based three-pole battery cell assembly equipment control method shows great potential in improving assembly accuracy and efficiency, it still faces some technical challenges: First, the shape and structure of the battery cells vary greatly, and highly robust image processing algorithms are needed to identify the different assembly states of the three-pole battery cells, which requires continuously optimized feature extraction and classification models. Secondly, the real-time and accuracy of real-time optical tracking technology are also challenges. The equipment must be able to quickly process a large amount of image data in a short period of time while maintaining positioning accuracy. In addition, the motion control and decision-making systems of the equipment need to be highly integrated to coordinate complex assembly actions and respond instantly to improve assembly efficiency. Summary of the invention

[0004] The present invention provides a control method and system for three-pole battery cell assembly equipment based on image recognition to solve the technical problems of low robustness and poor reliability of existing three-pole battery cell assembly equipment control methods and poor coordination and response ability of assembly equipment.

[0005] To achieve the above object, the present invention discloses a control method for three-electrode battery cell assembly equipment based on image recognition, comprising the following steps:

[0006] Acquire an actual assembly state image of the three-electrode battery cell at a preset time node, perform feature extraction processing on the actual assembly state image, and obtain a contour curve of the three-electrode battery cell;

[0007] The contour curve of the three-pole battery cell is discretized based on the grid method, and the discrete points obtained are subjected to redundancy correction. The actual assembly state model diagram of the three-pole battery cell is constructed according to the corrected discrete points.

[0008] Acquire preset assembly process information of the three-pole battery cell, and obtain a standard assembly state model diagram of the three-pole battery cell at a preset time node according to the preset assembly process information; calculate the degree of overlap between the actual assembly state model diagram and the standard assembly state model diagram based on a voxelization method;

[0009] The assembly state of the three-electrode battery cell is analyzed according to the degree of overlap between the actual assembly state model diagram and the standard assembly state model diagram, and the three-electrode battery cell assembly equipment is regulated according to the analysis result.

[0010] More specifically, feature extraction processing is performed on the actual assembly state image to obtain a contour curve of the three-electrode battery cell, specifically:

[0011] Binarization is performed on the actual assembly state image to simplify the value of each pixel in the actual assembly state image to 0 or 255; wherein the pixel simplified to 0 is defined as a black pixel, and the pixel simplified to 255 is defined as a white pixel;

[0012] In the actual assembly state image, the position nodes corresponding to the pixel points simplified to 0 are set to black, and the position nodes corresponding to the pixel points simplified to 255 are set to white, so as to obtain a binary image with only black and white;

[0013] Randomly select a black pixel point from the binary image as a tracking starting point, and traverse each black pixel point in the binary image from the tracking starting point;

[0014] During the traversal process, the neighboring points of each black pixel in the binary image in eight basic directions are retrieved based on the eight-direction chain code method;

[0015] Determine whether all neighboring points of each black pixel in the eight basic directions are black pixels; if all neighboring points of a black pixel in the eight basic directions are black pixels, mark the black pixel as a non-boundary point;

[0016] If the neighboring points of a black pixel in the eight basic directions are not all black pixels, the black pixel is marked as a boundary point;

[0017] If a black pixel is marked as a non-boundary point, the color of the position node corresponding to the black pixel in the binary image is reset to white; if a black pixel is marked as a boundary point, the color of the position node corresponding to the black pixel in the binary image is not modified;

[0018] This process is repeated in this way until all black pixels in the binary image are verified and corrected, and the contour curve of the three-electrode battery cell is obtained.

[0019] More specifically, the contour curve of the three-pole battery cell is discretized based on the grid method, and the discrete points obtained by discretization are subjected to redundancy correction processing, and the actual assembly state model diagram of the three-pole battery cell is constructed according to the corrected discrete points, specifically:

[0020] Setting the size and shape of the grid, constructing a grid model according to the set size and shape of the grid, and mapping the contour curve into the grid model;

[0021] Traverse each cell in the grid model and check whether each cell intersects with the contour curve; when a cell intersects with the contour curve, record the coordinates of the cell as a discrete point;

[0022] And so on, until all cells are traversed, several discrete points are obtained, and the coordinate information of each discrete point is obtained;

[0023] Calculate the Chebyshev distance between each discrete point according to the coordinate information of each discrete point; compare the Chebyshev distance between each discrete point with a preset distance threshold;

[0024] If the Chebyshev distance between two discrete points is not greater than the preset distance threshold, the two discrete points are marked as redundant discrete points;

[0025] If two discrete points are marked as redundant discrete points, any one of the discrete points is deleted; and so on, the process of performing redundant correction processing on the discrete points obtained by discretization is completed to obtain the corrected discrete points;

[0026] The coordinate information of the corrected discrete points is obtained, and a model diagram of the actual assembly state of the three-pole battery cell is constructed based on the coordinate information of the corrected discrete points and using three-dimensional software.

[0027] More specifically, the degree of overlap between the actual assembly state model diagram and the standard assembly state model diagram is calculated based on the voxelization method, specifically:

[0028] Creating a three-dimensional voxel grid, and dividing the three-dimensional voxel grid into a plurality of unit voxel grids; wherein the unit voxel grid is a cubic voxel grid, and the length, width and height of the unit voxel grid are all 1 mm;

[0029] Importing the actual assembly state model diagram and the standard assembly state model diagram into the three-dimensional voxel grid, and performing alignment processing on the assembly reference planes of the actual assembly state model diagram and the standard assembly state model diagram in the three-dimensional voxel grid;

[0030] After the alignment is completed, the actual assembly state model diagram and the standard assembly state model diagram are divided into a plurality of cubic voxel grids through each unit voxel grid; and each cubic voxel grid is traversed in a preset order;

[0031] If both the actual assembly state model diagram and the standard assembly state model diagram exist in a certain cubic voxel grid, the cubic voxel grid is calibrated as an overlapping voxel grid;

[0032] If only the actual assembly state model diagram or only the standard assembly state model diagram exists in a certain cubic voxel grid, the cubic voxel grid is calibrated as a non-overlapping voxel grid;

[0033] In the three-dimensional voxel grid, the cubic voxel grid area marked as the overlapping voxel grid is set to black, and the cubic voxel grid area marked as the non-overlapping voxel grid area is set to white;

[0034] The volume values ​​of the black area and the white area in the three-dimensional voxel grid are calculated, and the volume values ​​of the black area and the white area are ratio-processed to obtain the overlap degree between the actual assembly state model diagram and the standard assembly state model diagram.

[0035] More specifically, the assembly state of the three-pole battery cell is analyzed according to the overlap degree between the actual assembly state model diagram and the standard assembly state model diagram, and the three-pole battery cell assembly equipment is regulated according to the analysis result, specifically:

[0036] If the degree of overlap between the actual assembly state model diagram and the standard assembly state model diagram is greater than a preset overlap degree threshold, a first analysis result is generated, and the assembly equipment is controlled to execute a next preset assembly program;

[0037] If the degree of overlap between the actual assembly state model diagram and the standard assembly state model diagram is not greater than a preset overlap degree threshold, obtaining a region position of a white region in the three-dimensional voxel grid, defining the region position of the white region as an assembly offset region position, and obtaining an assembly offset region position of the three-pole battery cell;

[0038] Determine whether the assembly offset area of ​​the three-pole battery cell is an unrepairable area; if it is an unrepairable area, generate a second analysis result, and mark the three-pole battery cell as a scrap; if it is not an unrepairable area, generate a third analysis result, and mark the three-pole battery cell as a repairable product;

[0039] If the three-pole battery cell is marked as scrap, it will be transferred to the scrapping center; if the three-pole battery cell is marked as repairable, it will be transferred to the repair center.

[0040] The following steps are also included:

[0041] If the degree of overlap between the actual assembly state model diagram and the standard assembly state model diagram is not greater than a preset overlap degree threshold, obtaining the assembly offset area position;

[0042] Acquire preset assembly process information of the three-electrode battery cell, and determine the sub-assembly equipment that is assembly-related to the assembly offset area position according to the preset assembly process information;

[0043] Acquire the real-time working parameters of the subassembly equipment, compare the real-time working parameters with the preset working parameters, and obtain the working parameter deviation value;

[0044] It is determined whether the working parameter deviation value is greater than a preset threshold value; if greater than, the corresponding real-time working parameters of the sub-assembly equipment are regulated based on the working parameter deviation value.

[0045] To achieve the above-mentioned purpose, the present invention also discloses a three-pole battery cell assembly equipment control system based on image recognition, wherein the three-pole battery cell assembly equipment control system comprises a memory and a processor, wherein a three-pole battery cell assembly equipment control method program is stored in the memory, and when the three-pole battery cell assembly equipment control method program is executed by the processor, any step of the three-pole battery cell assembly equipment control method is implemented.

[0046] The present invention solves the technical defects existing in the background technology, and has the following beneficial effects: obtaining an actual assembly state image of a three-pole battery cell at a preset time node, performing feature extraction processing on the actual assembly state image, and obtaining a contour curve of the three-pole battery cell; discretizing the contour curve of the three-pole battery cell based on a gridding method, performing redundancy correction processing on the discrete points obtained by discretization, and constructing an actual assembly state model diagram of the three-pole battery cell according to the corrected discrete points; obtaining preset assembly process information of the three-pole battery cell, and obtaining a standard assembly state model diagram of the three-pole battery cell at a preset time node according to the preset assembly process information; calculating the degree of overlap between the actual assembly state model diagram and the standard assembly state model diagram based on a voxelization method; analyzing the assembly state of the three-pole battery cell according to the degree of overlap between the actual assembly state model diagram and the standard assembly state model diagram, and regulating the three-pole battery cell assembly equipment according to the analysis result. The assembly control method based on image recognition can significantly improve assembly accuracy, reduce scrap rate, and ensure product quality while also improving production efficiency. Through real-time feedback and automatic correction, it reduces the need for manual intervention and realizes intelligent and refined control of the production process. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, drawings of other embodiments can be obtained based on these drawings without paying creative work.

[0048] Figure 1 A first method flow chart of a method for controlling a three-electrode battery cell assembly device based on image recognition;

[0049] Figure 2 A second method flow chart of a method for controlling a three-electrode battery cell assembly device based on image recognition;

[0050] Figure 3 This is the system block diagram of the control system of the three-pole battery cell assembly equipment based on image recognition. DETAILED DESCRIPTION

[0051] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0052] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0053] like Figure 1 As shown, the present invention discloses a control method for three-electrode battery cell assembly equipment based on image recognition, comprising the following steps:

[0054] S102: acquiring an actual assembly state image of the three-tab battery cell at a preset time node, performing feature extraction processing on the actual assembly state image, and obtaining a contour curve of the three-tab battery cell;

[0055] S104: discretizing the contour curve of the three-electrode battery cell based on a gridding method, performing redundancy correction processing on the discrete points obtained by discretization, and constructing an actual assembly state model diagram of the three-electrode battery cell according to the corrected discrete points;

[0056] S106: Acquire preset assembly process information of the three-pole battery cell, and obtain a standard assembly state model diagram of the three-pole battery cell at a preset time node according to the preset assembly process information; and calculate the degree of overlap between the actual assembly state model diagram and the standard assembly state model diagram based on a voxelization method;

[0057] S108: analyzing the assembly state of the three-electrode battery cell according to the degree of overlap between the actual assembly state model diagram and the standard assembly state model diagram, and regulating the three-electrode battery cell assembly equipment according to the analysis result.

[0058] Among them, the preset assembly process information of three-electrode battery cells refers to a set of pre-set detailed guidelines, which include standardized operation procedures, parameter settings, component configurations and equipment usage methods for each link in the battery cell manufacturing process, such as assembly steps, process parameters, component specifications, quality control indicators, etc.

[0059] Among them, the standard assembly state model diagram of the three-pole battery cell is an idealized three-dimensional model that shows the expected state of the battery cell during the assembly process under ideal conditions. This model usually includes the precise layout of the various components of the battery cell (such as the tabs, electrode sheets, casing, etc.) in the correct assembly position, as well as the geometric relationship and gap between them. It is created according to the design specifications and standard process flow of the battery cell and is used as a reference benchmark in the production process. In the actual assembly process, by comparing the degree of overlap between the actual assembly state model diagram and the standard model diagram, it is possible to detect whether there is a deviation, so as to adjust the process parameters or equipment operation in time to achieve matching with the standard and improve the accuracy and consistency of assembly.

[0060] More specifically, feature extraction processing is performed on the actual assembly state image to obtain a contour curve of the three-electrode battery cell, specifically:

[0061] Binarization is performed on the actual assembly state image to simplify the value of each pixel in the actual assembly state image to 0 or 255; wherein the pixel simplified to 0 is defined as a black pixel, and the pixel simplified to 255 is defined as a white pixel;

[0062] In the actual assembly state image, the position nodes corresponding to the pixel points simplified to 0 are set to black, and the position nodes corresponding to the pixel points simplified to 255 are set to white, so as to obtain a binary image with only black and white;

[0063] Randomly select a black pixel point from the binary image as a tracking starting point, and traverse each black pixel point in the binary image from the tracking starting point;

[0064] During the traversal process, the neighboring points of each black pixel in the binary image in eight basic directions are retrieved based on the eight-direction chain code method;

[0065] Determine whether all neighboring points of each black pixel in the eight basic directions are black pixels; if all neighboring points of a black pixel in the eight basic directions are black pixels, mark the black pixel as a non-boundary point;

[0066] If the neighboring points of a black pixel in the eight basic directions are not all black pixels, the black pixel is marked as a boundary point;

[0067] If a black pixel is marked as a non-boundary point, the color of the position node corresponding to the black pixel in the binary image is reset to white; if a black pixel is marked as a boundary point, the color of the position node corresponding to the black pixel in the binary image is not modified;

[0068] This process is repeated in this way until all black pixels in the binary image are verified and corrected, and the contour curve of the three-electrode battery cell is obtained.

[0069] It should be noted that the industrial camera is controlled to obtain the actual assembly state image of the three-pole battery cell at a preset time node, such as the actual assembly state image of the three-pole battery cell after the electrode sheet is assembled. Then the actual assembly image is binarized, and the pixel values ​​in the image are simplified to 0 (black) or 255 (white) to facilitate the identification of the assembly state. By converting the image into a black and white binary image, the boundary information of the battery cell assembly can be clearly displayed. Then, starting from a randomly selected black pixel, the eight-directional chain code method is used to traverse all black pixels in the image, and at the same time check whether its adjacent pixels are also black. In this way, the algorithm can identify boundary points (points surrounded by non-black pixels) and non-boundary points (points surrounded by all black). For non-boundary points, the algorithm changes the color of these points from black to white, thereby eliminating the interference of boundary points on contour extraction. For boundary points, the algorithm keeps them black because they are the key part of the battery cell contour. This process continues until all black pixels are processed, and the final result is a clearer and more accurate contour curve of the three-pole battery cell, which enables the subsequent construction of a more realistic model diagram of the actual assembly state of the three-pole battery cell. This can significantly improve the visualization and analysis capabilities of the assembly state, accurately identify possible problems in the assembly process, and help improve assembly quality and production efficiency.

[0070] More specifically, the contour curve of the three-pole battery cell is discretized based on the grid method, and the discrete points obtained by discretization are subjected to redundancy correction processing, and the actual assembly state model diagram of the three-pole battery cell is constructed according to the corrected discrete points, specifically:

[0071] Setting the size and shape of the grid, constructing a grid model according to the set size and shape of the grid, and mapping the contour curve into the grid model;

[0072] Traverse each cell in the grid model and check whether each cell intersects with the contour curve; when a cell intersects with the contour curve, record the coordinates of the cell as a discrete point;

[0073] And so on, until all cells are traversed, several discrete points are obtained, and the coordinate information of each discrete point is obtained;

[0074] Calculate the Chebyshev distance between each discrete point according to the coordinate information of each discrete point; compare the Chebyshev distance between each discrete point with a preset distance threshold;

[0075] If the Chebyshev distance between two discrete points is not greater than the preset distance threshold, the two discrete points are marked as redundant discrete points;

[0076] If two discrete points are marked as redundant discrete points, any one of the discrete points is deleted; and so on, the process of performing redundant correction processing on the discrete points obtained by discretization is completed to obtain the corrected discrete points;

[0077] The coordinate information of the corrected discrete points is obtained, and a model diagram of the actual assembly state of the three-pole battery cell is constructed based on the coordinate information of the corrected discrete points and using three-dimensional software.

[0078] It should be noted that the gridding method is a method of discretizing a continuous curve. It creates a grid in the area where the curve is located, and then finds the intersection of the curve and the grid to divide the curve into a series of discrete points. The size and shape of the grid are set in advance. Usually, the size of the grid is determined by the required discrete point accuracy. The grid can be square, rectangular or other shapes.

[0079] By setting the size and shape of the grid, the contour curve is mapped to the grid model, and the intersection of each cell and the contour is regarded as a discrete point. Then, each cell in the grid is traversed to record the coordinates of the points that intersect with the curve. The Chebyshev distance is used to calculate the closeness between adjacent discrete points. If the distance is less than the preset threshold, it is marked as a redundant point and one is deleted. Finally, according to the coordinate information of the corrected discrete points and using three-dimensional software (such as proe, cad, etc.), the actual assembly state model diagram of the three-pole ear battery cell is constructed to facilitate subsequent analysis and optimization. This method can improve the accuracy of discrete points and the accuracy of the model, reduce data redundancy by eliminating redundancy, help reduce the computational complexity in the subsequent modeling process, improve computing efficiency, and the final assembly state model diagram can more realistically reflect the actual assembly of the battery cell, which is conducive to quality control and optimization of the assembly process.

[0080] In addition, it should be noted that the discrete points obtained by the gridding method may be redundant. This is because the gridding process usually detects the intersection of the curve based on the boundary of the grid cells. For example, if the curve is exactly tangent to or intersects with the boundary of the grid, there may be a discrete point on each side of the boundary. This is actually a different representation of the same point, resulting in redundant points.

[0081] In general, the advantage of the gridding method is that it is simple and intuitive, but the disadvantage is that more discrete points may be generated during the discretization process, especially when the curves are dense or the grids are fine. Therefore, in this method, it is necessary to correct the redundant points obtained by discretization to reduce the amount of data and improve the calculation efficiency.

[0082] More specifically, the degree of overlap between the actual assembly state model diagram and the standard assembly state model diagram is calculated based on the voxelization method, specifically:

[0083] Creating a three-dimensional voxel grid, and dividing the three-dimensional voxel grid into a plurality of unit voxel grids; wherein the unit voxel grid is a cubic voxel grid, and the length, width and height of the unit voxel grid are all 1 mm;

[0084] Importing the actual assembly state model diagram and the standard assembly state model diagram into the three-dimensional voxel grid, and performing alignment processing on the assembly reference planes of the actual assembly state model diagram and the standard assembly state model diagram in the three-dimensional voxel grid;

[0085] After the alignment is completed, the actual assembly state model diagram and the standard assembly state model diagram are divided into a plurality of cubic voxel grids through each unit voxel grid; and each cubic voxel grid is traversed in a preset order;

[0086] If both the actual assembly state model diagram and the standard assembly state model diagram exist in a certain cubic voxel grid, the cubic voxel grid is calibrated as an overlapping voxel grid;

[0087] If only the actual assembly state model diagram or only the standard assembly state model diagram exists in a certain cubic voxel grid, the cubic voxel grid is calibrated as a non-overlapping voxel grid;

[0088] In the three-dimensional voxel grid, the cubic voxel grid area marked as the overlapping voxel grid is set to black, and the cubic voxel grid area marked as the non-overlapping voxel grid area is set to white;

[0089] The volume values ​​of the black area and the white area in the three-dimensional voxel grid are calculated, and the volume values ​​of the black area and the white area are ratio-processed to obtain the overlap degree between the actual assembly state model diagram and the standard assembly state model diagram.

[0090] It should be noted that, first, by creating a cubic voxel grid with a side length of 1 mm, the actual and standard assembly model drawings are imported and aligned respectively to ensure that the two are on the same assembly reference plane. Then, according to the voxel grid, the model is divided into overlapping and non-overlapping areas, black represents overlapping areas, and white represents non-overlapping areas. Next, by calculating the volume of the black and white areas, their volume ratio is obtained to quantitatively evaluate the degree of overlap between the actual assembly state model drawing and the standard model drawing. This method can accurately quantify the degree of match between the actual assembly state model drawing and the standard model drawing, which is suitable for large-scale three-dimensional model comparison, and can quickly determine the deviation or consistency of the three-pole ear battery during the assembly process. In this way, the differences or problem areas in the assembly can be intuitively identified, and the movement and decision-making of the assembly equipment can be highly integrated and controlled to coordinate complex assembly actions and respond immediately to improve assembly efficiency.

[0091] More specifically, if Figure 2 As shown, the assembly state of the three-pole battery cell is analyzed according to the overlap degree between the actual assembly state model diagram and the standard assembly state model diagram, and the three-pole battery cell assembly equipment is regulated according to the analysis result, specifically:

[0092] S202: If the overlap degree between the actual assembly state model diagram and the standard assembly state model diagram is greater than a preset overlap degree threshold, a first analysis result is generated, and the assembly equipment is controlled to execute a next preset assembly program;

[0093] S204: if the degree of overlap between the actual assembly state model diagram and the standard assembly state model diagram is not greater than a preset overlap degree threshold, obtaining a region position of a white region in the three-dimensional voxel grid, defining the region position of the white region as an assembly offset region position, and obtaining an assembly offset region position of the three-electrode battery cell;

[0094] S206: Determine whether the assembly offset region of the three-pole battery cell is an unrepairable region; if it is an unrepairable region, generate a second analysis result, and mark the three-pole battery cell as a waste; if it is not an unrepairable region, generate a third analysis result, and mark the three-pole battery cell as a repairable product;

[0095] S208: If the three-pole lug battery cell is marked as scrap, the three-pole lug battery cell is transferred to a scrapping center; if the three-pole lug battery cell is marked as repairable, the three-pole lug battery cell is transferred to a repair center.

[0096] Among them, during the assembly process, if the deviation of the main structure of the battery cell (such as the tab, electrode sheet or shell) causes damage to the structural integrity, such as tab breakage, electrode sheet misalignment, or shell deformation, these will seriously affect the function of the battery cell; for example, scratches, dents or corrosion on the surface of the battery cell, these defects will affect the appearance quality of the battery cell, and will also lead to exposure of internal materials, resulting in performance degradation. This type of damage usually cannot be restored by repair; for example, the connection between the electrode and the negative tab, and the positive tab and the battery pack, if these connections are seriously offset, it will lead to poor contact, affecting current transmission, and it is impossible to restore good electrical performance through simple adjustment; when the above areas are offset in assembly, they are all irreparable areas.

[0097] It should be noted that first, by comparing the degree of overlap between the actual assembly state model diagram and the standard model diagram, if it exceeds the preset threshold, it indicates that the assembly is normal, the system generates a first analysis result, and controls the assembly equipment to execute the next preset assembly procedure. If the overlap is insufficient, the system will identify the assembly offset area, that is, the position of the white area, as a deviation reference for the battery cell assembly. Next, the system will determine whether this offset area belongs to an unrepairable area. If this offset area belongs to an unrepairable area, the battery cell will be marked as unqualified, that is, scrap, and then transferred to the scrap center. On the contrary, if this offset area does not belong to an unrepairable area, the system generates a third analysis result, marks it as a repairable product, and sends it to the repair center for adjustment. This method improves the quality control efficiency of the assembly process, reduces manual intervention through automated analysis, reduces the error rate, and can quickly locate the problem area, effectively manage product quality, and classify scrap and repairable products, which helps to optimize resource allocation and reduce costs.

[0098] The following steps are also included:

[0099] If the degree of overlap between the actual assembly state model diagram and the standard assembly state model diagram is not greater than a preset overlap degree threshold, obtaining the assembly offset area position;

[0100] Acquire preset assembly process information of the three-electrode battery cell, and determine the sub-assembly equipment that is assembly-related to the assembly offset area position according to the preset assembly process information;

[0101] Acquire the real-time working parameters of the subassembly equipment, compare the real-time working parameters with the preset working parameters, and obtain the working parameter deviation value;

[0102] It is determined whether the working parameter deviation value is greater than a preset threshold value; if greater than, the corresponding real-time working parameters of the sub-assembly equipment are regulated based on the working parameter deviation value.

[0103] It should be noted that by comparing the degree of overlap between the actual assembly state model diagram and the standard model diagram, if an offset is found, the system will locate the specific assembly offset area. Then, the system determines the sub-assembly equipment related to the offset area based on the preset assembly process information. Then, the system monitors the real-time working parameters of the sub-assembly equipment, such as speed, pressure, etc., compares it with the preset parameters, and calculates the deviation value. If the deviation value exceeds the preset threshold, the system will immediately judge it as abnormal and adjust the working parameters of the sub-assembly equipment to correct the problem in the assembly process. This method can significantly improve the assembly accuracy, reduce the scrap rate, ensure product quality, and improve production efficiency. Through real-time feedback and automatic deviation correction, the need for manual intervention is reduced, and intelligent and refined control of the production process is realized.

[0104] The preset assembly process information includes detailed assembly steps, the relationship between components, and the equipment and operation sequence required for each step. This information includes: (1) Assembly step association. Each assembly step has specific equipment and actions, such as installing electrode ears and assembling shells. By matching the preset steps, it is possible to determine which equipment is responsible for assembly in a specific area. (2) Component positioning. The preset process information contains the precise location and dependency of the components during the assembly process. If assembly offset occurs at a specific location, the subassembly equipment near that location may be related to it. (3) Equipment responsibility area. Each subassembly equipment may be assigned a specific working area, such as the electrode ear installer may be responsible for electrode installation, while the shell closing machine may be responsible for shell matching. If the offset area is within the operating range of these equipment, it indicates that the equipment may be related to the offset. (4) It includes settings for equipment working parameters (such as pressure, speed, accuracy, etc.). The degree of matching of these parameters with the assembly area can also indicate the relevance of the equipment. By comprehensively analyzing this information, the system can infer which subassembly equipment may cause problems due to improper process parameters or incorrect assembly sequence when assembly offset occurs, and thus take corresponding control measures. This not only improves assembly accuracy, but also simplifies the troubleshooting process.

[0105] In addition, the method further comprises the following steps:

[0106] Obtaining operation log information of the assembly equipment, obtaining fault events of each sub-assembly equipment according to the operation log information, and performing feature engineering processing on the fault events of each sub-assembly equipment to obtain a fault feature data set of each fault event occurring in each sub-assembly equipment;

[0107] Constructing a knowledge graph, and importing the fault feature data set of each fault event occurring in each sub-assembly device into the knowledge graph;

[0108] Obtaining the assembly offset area position of each three-pole battery cell when the assembly equipment assembles the three-pole battery cell within a preset time period;

[0109] The three-pole battery cell is divided into a plurality of sub-assembly areas, and according to the assembly offset area position of each three-pole battery cell when the assembly equipment assembles the three-pole battery cell within a preset time period, the assembly abnormality probability of each sub-assembly area of ​​the assembly equipment during the assembly process is counted;

[0110] If the probability of assembly anomaly in a certain sub-assembly area is greater than the preset probability value, the sub-assembly area is marked as a frequent assembly deviation area; if the probability of assembly anomaly in a certain sub-assembly area is not greater than the preset probability value, the sub-assembly area is marked as an occasional assembly deviation area;

[0111] If a sub-assembly area is marked as an assembly offset accidental area, then the sub-assembly equipment that is assembly-related to the sub-assembly area is obtained and defined as suspicious sub-assembly equipment;

[0112] Acquire an actual working parameter set of the suspicious subassembly equipment within a second preset time period, import the actual working parameter set into the knowledge graph, and calculate the similarity between the actual working parameter set and each fault feature data set based on a Euclidean distance algorithm;

[0113] If the similarities between the actual working parameter set and each fault feature data set are not greater than the preset similarities, the suspicious subassembly equipment is continuously monitored;

[0114] If the similarity between the actual working parameter set and a certain fault feature data set is greater than the preset similarity, equipment fault information is generated, and the equipment fault type is obtained according to the corresponding fault event, and the equipment fault information and the equipment fault type are sent to the preset platform for display; and the assembly equipment is controlled to stop production.

[0115] It should be noted that, first, the system collects the log information of the assembly equipment, extracts the features of the fault events of the sub-assembly equipment, and forms a fault feature data set. Then, by constructing a knowledge graph, these data are associated with the equipment operation mode. During the assembly process, the system monitors the assembly offset of the three-pole battery cell and determines whether it is a frequent or accidental offset based on the abnormal probability. For areas with frequent offsets, the system will mark the relevant sub-assembly equipment as suspicious objects. In the second preset time period, the system will further collect the working parameters of the suspicious equipment and compare their similarity with the fault feature data set through the Euclidean distance algorithm. If the working parameters are highly matched with the fault characteristics, the system will identify the specific fault type, generate fault information and send it to the preset platform, and trigger the equipment to shut down for maintenance. The method can effectively improve the maintenance efficiency and accuracy of the assembly equipment, prevent and locate faults in advance through data analysis, thereby ensuring the stable operation of the production line; and realize the function of fault self-diagnosis, which can avoid the situation of batch production of defective products due to equipment failure and reduce scrap costs.

[0116] In addition, the method further comprises the following steps:

[0117] If it is determined that the assembly equipment fails, the equipment failure information of the assembly equipment is obtained, and the predicted maintenance time required for repairing the assembly equipment is retrieved through the big data network according to the equipment failure information;

[0118] Obtaining average production capacity information of the assembly equipment, and calculating the reduction in production when repairing the faulty assembly equipment based on the average production capacity information and the predicted repair time;

[0119] Obtain the idle inventory of three-electrode battery cells in the production workshop, and determine whether the idle inventory is greater than the reduced production;

[0120] If it is not greater than, searching for assembly equipment of the same type as the faulty assembly equipment in the production workshop, and obtaining the status information of the assembly equipment of the same type as the faulty assembly equipment, and defining the assembly equipment whose status information is idle as the movable assembly equipment;

[0121] The assembly accuracy of each adjustable assembly device is obtained, as well as the assembly accuracy of the faulty assembly device, and the adjustable assembly device with the highest assembly accuracy matching degree is output as the recommended energy replenishment device.

[0122] It should be noted that once a failure of an assembly device is detected, the system first obtains the failure information of the device and uses the big data network to retrieve relevant maintenance history and cases to predict the time required for maintenance. Then, the system calculates the possible reduction in production during the maintenance period based on the average production capacity of the assembly equipment and the predicted maintenance time. In order to cope with possible production gaps, the system will also query the inventory of three-pole battery cells in the production workshop to determine whether the current idle inventory is sufficient to make up for the reduction in production caused by maintenance. If the inventory is not enough to cover the reduction in production, the system will further search for assembly equipment of the same type and the current status is "idle", which are considered as potential alternative options. The system will evaluate the assembly accuracy of each movable assembly equipment and compare it with the assembly accuracy of the faulty equipment to find the equipment with the highest matching degree of assembly accuracy as the recommended alternative equipment. Through such a process, the system can not only respond to equipment failures quickly and predict maintenance time, but also effectively manage production plans, ensure the continuity and efficiency of production lines, help reduce production losses caused by equipment downtime, and improve overall production flexibility and response speed. At the same time, by optimizing resource allocation and equipment use, production efficiency and product quality can also be improved to achieve more sustainable production operations.

[0123] like Figure 3 As shown, the present invention also discloses a three-pole battery cell assembly equipment control system based on image recognition, the three-pole battery cell assembly equipment control system includes a memory 41 and a processor 62, the memory 41 stores a three-pole battery cell assembly equipment control method program, when the three-pole battery cell assembly equipment control method program is executed by the processor 62, any step of the three-pole battery cell assembly equipment control method is implemented.

[0124] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0125] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0126] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0127] A person of ordinary skill in the art can understand that: all or part of the steps of implementing the above method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiment; and the aforementioned storage medium includes: a mobile storage device, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0128] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the methods of each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

[0129] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art who is familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A control method for three-electrode battery cell assembly equipment based on image recognition, characterized in that: The following steps are involved: Acquire an actual assembly state image of the three-electrode battery cell at a preset time node, perform feature extraction processing on the actual assembly state image, and obtain a contour curve of the three-electrode battery cell; The contour curve of the three-pole battery cell is discretized based on the grid method, and the discrete points obtained are subjected to redundancy correction. The actual assembly state model diagram of the three-pole battery cell is constructed according to the corrected discrete points. Acquire preset assembly process information of the three-pole battery cell, and obtain a standard assembly state model diagram of the three-pole battery cell at a preset time node according to the preset assembly process information; Calculating the degree of overlap between the actual assembly state model diagram and the standard assembly state model diagram based on a voxelization method; Analyzing the assembly state of the three-electrode battery cell according to the degree of overlap between the actual assembly state model diagram and the standard assembly state model diagram, and regulating the three-electrode battery cell assembly equipment according to the analysis result; Among them, the contour curve of the three-pole battery cell is discretized based on the grid method, and the discrete points obtained by discretization are subjected to redundant correction processing. According to the corrected discrete points, the actual assembly state model diagram of the three-pole battery cell is constructed, which is specifically: Setting the size and shape of the grid, constructing a grid model according to the set size and shape of the grid, and mapping the contour curve into the grid model; Traverse each cell in the grid model and check whether each cell intersects with the contour curve; when a cell intersects with the contour curve, record the coordinates of the cell as a discrete point; And so on, until all cells are traversed, several discrete points are obtained, and the coordinate information of each discrete point is obtained; Calculate the Chebyshev distance between each discrete point according to the coordinate information of each discrete point; compare the Chebyshev distance between each discrete point with a preset distance threshold; If the Chebyshev distance between two discrete points is not greater than the preset distance threshold, the two discrete points are marked as redundant discrete points; If two discrete points are marked as redundant discrete points, any one of them is deleted; and so on, the process of performing redundant correction processing on the discrete points obtained by discretization is completed to obtain the corrected discrete points; The coordinate information of the corrected discrete points is obtained, and a model diagram of the actual assembly state of the three-pole battery cell is constructed based on the coordinate information of the corrected discrete points and using three-dimensional software.

2. The method for controlling a three-electrode battery cell assembly device based on image recognition according to claim 1, characterized in that: The actual assembly state image is subjected to feature extraction processing to obtain a contour curve of the three-electrode battery cell, specifically: Binarization is performed on the actual assembly state image to simplify the value of each pixel in the actual assembly state image to 0 or 255; wherein the pixel simplified to 0 is defined as a black pixel, and the pixel simplified to 255 is defined as a white pixel; In the actual assembly state image, the position nodes corresponding to the pixel points simplified to 0 are set to black, and the position nodes corresponding to the pixel points simplified to 255 are set to white, so as to obtain a binary image with only black and white; Randomly select a black pixel point from the binary image as a tracking starting point, and traverse each black pixel point in the binary image from the tracking starting point; During the traversal process, the neighboring points of each black pixel in the binary image in eight basic directions are retrieved based on the eight-direction chain code method; Determine whether all neighboring points of each black pixel in the eight basic directions are black pixels; if all neighboring points of a black pixel in the eight basic directions are black pixels, mark the black pixel as a non-boundary point; If the neighboring points of a black pixel in the eight basic directions are not all black pixels, the black pixel is marked as a boundary point; If a black pixel is marked as a non-boundary point, the color of the position node corresponding to the black pixel in the binary image is reset to white; if a black pixel is marked as a boundary point, the color of the position node corresponding to the black pixel in the binary image is not modified; This process is repeated in this way until all black pixels in the binary image are verified and corrected, and the contour curve of the three-electrode battery cell is obtained.

3. The method for controlling a three-electrode battery cell assembly device based on image recognition according to claim 1, characterized in that: The degree of overlap between the actual assembly state model diagram and the standard assembly state model diagram is calculated based on the voxelization method, specifically: Creating a three-dimensional voxel grid, and dividing the three-dimensional voxel grid into a plurality of unit voxel grids; wherein the unit voxel grid is a cubic voxel grid, and the length, width and height of the unit voxel grid are all 1 mm; Importing the actual assembly state model diagram and the standard assembly state model diagram into the three-dimensional voxel grid, and performing alignment processing on the assembly reference planes of the actual assembly state model diagram and the standard assembly state model diagram in the three-dimensional voxel grid; After the alignment is completed, the actual assembly state model diagram and the standard assembly state model diagram are divided into a plurality of cubic voxel grids through each unit voxel grid; and each cubic voxel grid is traversed in a preset order; If both the actual assembly state model diagram and the standard assembly state model diagram exist in a certain cubic voxel grid, the cubic voxel grid is calibrated as an overlapping voxel grid; If only the actual assembly state model diagram or only the standard assembly state model diagram exists in a certain cubic voxel grid, the cubic voxel grid is calibrated as a non-overlapping voxel grid; In the three-dimensional voxel grid, the cubic voxel grid area marked as the overlapping voxel grid is set to black, and the cubic voxel grid area marked as the non-overlapping voxel grid area is set to white; The volume values ​​of the black area and the white area in the three-dimensional voxel grid are calculated, and the volume values ​​of the black area and the white area are ratio-processed to obtain the overlap degree between the actual assembly state model diagram and the standard assembly state model diagram.

4. The method for controlling a three-electrode battery cell assembly device based on image recognition according to claim 3, characterized in that: The assembly state of the three-electrode battery cell is analyzed according to the overlap degree between the actual assembly state model diagram and the standard assembly state model diagram, and the three-electrode battery cell assembly equipment is regulated according to the analysis result, specifically: If the degree of overlap between the actual assembly state model diagram and the standard assembly state model diagram is greater than a preset overlap degree threshold, a first analysis result is generated, and the assembly equipment is controlled to execute a next preset assembly program; If the degree of overlap between the actual assembly state model diagram and the standard assembly state model diagram is not greater than a preset overlap degree threshold, obtaining a region position of a white region in the three-dimensional voxel grid, defining the region position of the white region as an assembly offset region position, and obtaining an assembly offset region position of the three-pole battery cell; Determine whether the assembly offset area of ​​the three-electrode battery cell is an unrepairable area; If it is an unrepairable area, a second analysis result is generated, and the three-electrode battery cell is marked as waste; If it is not an unrepairable area, a third analysis result is generated, and the three-electrode battery cell is marked as a repairable product; If the three-pole battery cell is marked as scrap, it will be transferred to the scrapping center; if the three-pole battery cell is marked as repairable, it will be transferred to the repair center.

5. The method for controlling a three-electrode battery cell assembly device based on image recognition according to claim 4, characterized in that: The following steps are also included: If the degree of overlap between the actual assembly state model diagram and the standard assembly state model diagram is not greater than a preset overlap degree threshold, obtaining the assembly offset area position; Acquire preset assembly process information of the three-electrode battery cell, and determine the sub-assembly equipment that is assembly-related to the assembly offset area position according to the preset assembly process information; Acquire the real-time working parameters of the subassembly equipment, compare the real-time working parameters with the preset working parameters, and obtain the working parameter deviation value; It is determined whether the working parameter deviation value is greater than a preset threshold value; if greater than, the corresponding real-time working parameters of the sub-assembly equipment are regulated based on the working parameter deviation value.

6. A control system for three-electrode battery cell assembly equipment based on image recognition, characterized in that: The three-pole battery cell assembly equipment control system includes a memory and a processor, wherein the memory stores a three-pole battery cell assembly equipment control method program, and when the three-pole battery cell assembly equipment control method program is executed by the processor, the steps of the three-pole battery cell assembly equipment control method as described in any one of claims 1 to 5 are implemented.

Images