Battery pack debonding detection methods, apparatus, equipment, storage media, and procedures.

Abstract

This application relates to a method, apparatus, device, storage medium, and program product for detecting battery pack delamination. The method includes: performing target detection on an initial image of the battery pack using a pre-trained position detection model to obtain the position information of each cell in the battery pack; wherein, in the initial image, each cell is arranged in its original position before disassembly and has been impregnated with a color developer; performing region segmentation based on the position information and the impregnation color of the color developer to determine the bottom surface image of the delamination cell, and performing semantic analysis based on the impregnation color of the color developer to identify marking information within the bottom surface image; and determining the delamination detection result of the battery pack based on the marking information within the bottom surface image. Using this application can improve detection efficiency and increase the detection accuracy of battery pack delamination.

Description

Technical Field

[0001] This application relates to the field of battery testing technology, specifically to a method, apparatus, equipment, storage medium, and program product for detecting battery pack delamination. Background Technology

[0002] With the widespread application of lithium-ion batteries in new energy vehicles, energy storage systems, and other fields, the reliability and safety of battery packs have become a key concern. In practical use, battery packs may be subjected to vibration and impact, causing the cells to delaminate against the base plate, thus affecting the performance and lifespan of the battery pack.

[0003] Currently, the method for detecting delamination uses red ink as a colorimetric agent, and the area of ​​delamination is determined manually by visual inspection after the battery pack is disassembled. However, this method suffers from low detection efficiency and insufficient detection accuracy. Summary of the Invention

[0004] To address the aforementioned issues, this application provides a method, apparatus, equipment, storage medium, and program product for detecting battery pack delamination, which can meet the needs of large-scale testing, improve testing efficiency, and enhance the accuracy of battery pack delamination detection.

[0005] In a first aspect, this application provides a method for detecting battery pack delamination, used to detect whether delamination has occurred at the bottom of the battery cells in the battery pack, the method comprising:

[0006] The initial image of the battery pack is used to perform target detection using a pre-trained position detection model to obtain the position information of each cell in the battery pack; wherein, in the initial image, each cell is arranged in its original position before the battery pack is disassembled and has been impregnated with color developer;

[0007] Region segmentation is performed based on location information and the wetting color of the color developer to determine the bottom image of the debonded battery cell, and semantic analysis is performed based on the wetting color of the color developer to identify the marking information in the bottom image.

[0008] Based on the marking information in the bottom image, the debonding detection result of the battery pack is determined.

[0009] In the technical solution of this application embodiment, using images for automatic debonding detection can meet the needs of large-scale detection and improve detection efficiency. Furthermore, this solution first accurately locates the disassembled battery cells, which can reduce the problem of inaccurate detection caused by factors such as shooting posture and irregular battery cell arrangement. Based on the location, the bottom surface image is extracted, which can further improve the detection accuracy of the debonded area, thereby improving the detection accuracy of battery pack debonding.

[0010] In some embodiments, a pre-trained location detection model is used to perform target detection on an initial image of the battery pack to obtain the location information of each cell in the battery pack, including:

[0011] The initial image is input into the location detection model for target detection, and the corner coordinates of multiple battery cells are obtained from the output of the location detection model.

[0012] The corner coordinates of multiple battery cells are mapped, and the position information of each battery cell in the battery pack is determined based on the mapping results.

[0013] In the technical solution of this application embodiment, the use of position detection model and two-dimensional mesh mapping processing can reduce the influence of shooting posture and cell arrangement, reduce the dependence of shooting on tooling fixtures, and improve the robustness of the overall detection process.

[0014] In some embodiments, region segmentation is performed based on location information and the wetting color of the color developer to determine the bottom surface image of the debonded battery cell, including:

[0015] Based on the location information of each cell in the battery pack, the location of each cell is determined in the initial image;

[0016] The area of ​​each battery cell is tested based on the wetting color of the color developer to obtain preliminary test results for each battery cell; the preliminary test results are used to characterize whether the battery cell has delamination.

[0017] Based on the preliminary detection results, the initial image was segmented to determine the bottom image of the debonded battery cell.

[0018] In the technical solution of this application embodiment, the debonded battery cells can be selected based on the wetting color of the color developer, so that only the debonded battery cells are tested in the subsequent process, and the non-debonded battery cells are not tested, thereby improving the testing efficiency.

[0019] In some embodiments, the initial image is segmented based on preliminary detection results to determine the bottom image of the debonded battery cell, including:

[0020] Based on the preliminary test results, the initial cell image was cropped to obtain a regional image of the area where the debonded cell was located.

[0021] The bottom surface image of the debonded battery cell is obtained by using a pre-trained bottom surface recognition model to identify and process the regional image.

[0022] In the technical solution of this application embodiment, bottom surface recognition is performed using the cropped region image, which can reduce the computational burden and training difficulty of the bottom surface recognition model.

[0023] In some embodiments, a pre-trained bottom surface recognition model is used to process the region image to obtain a bottom surface image of the debonded battery cell, including:

[0024] The region image is input into the bottom surface recognition model for recognition processing to obtain the bottom surface location of the debonded battery cell;

[0025] The image of the debonded battery cell is cropped based on the location of its bottom surface to obtain the bottom surface image of the debonded battery cell.

[0026] In the technical solution of this application embodiment, the bottom surface recognition model is used for bottom surface recognition, which can reduce the impact of image deformation on recognition accuracy, thereby enhancing the accuracy of bottom surface image extraction. It is especially suitable for soft-pack batteries or irregularly shaped batteries. Even if the battery cell has an irregular shape or slight adhesion, the bottom surface image can be effectively separated.

[0027] In some embodiments, semantic analysis is performed based on the wetting color of the developer to identify marker information within the bottom image, including:

[0028] For each debonded battery cell, the bottom image is color-converted to obtain the converted image;

[0029] Semantic analysis is performed on the converted image based on the wetting color of the developer to identify the marking information in the bottom image.

[0030] In the technical solution of this application embodiment, the bottom area of ​​the debonded battery cell is color-converted, which is beneficial for color semantic analysis and can improve detection accuracy.

[0031] In some embodiments, semantic analysis is performed on the converted image based on the wetting color of the developer to identify marker information within the bottom image, including:

[0032] The initial degummed area in the converted image is obtained by identifying and processing the wetting color of the developer.

[0033] The wetting characteristics of the initial degummed area are determined, and the marking information of the bottom image is obtained by filling the wetting characteristics.

[0034] In the technical solution of this application embodiment, the degummed area is reconstructed, which improves the completeness of identification of weak or localized impregnation and reduces the risk of missed detection.

[0035] In some embodiments, the marking information of the bottom image is obtained by filling based on the infiltration features, including:

[0036] When the wetting feature is that the wetting area does not include the edge of the battery cell and forms a closed loop at the center of the bottom surface, the outer ring at the center of the bottom surface is filled to obtain the marking information of the bottom surface image.

[0037] In some embodiments, the marking information of the bottom image is obtained by filling based on the infiltration features, including:

[0038] When the wetting feature is that the wetting area includes the edge of the battery cell but there are gaps, the area between the center of the bottom surface and the edge of the battery cell is filled to obtain the marking information of the bottom surface image;

[0039] In some embodiments, the marking information of the bottom image is obtained by filling based on the infiltration features, including:

[0040] When the wetting feature is that the wetting area does not include the edge of the battery cell and no closed loop is formed at the center of the bottom surface, the unclosed position at the center of the bottom surface is closed, and the outer ring of the closed center of the bottom surface is filled to obtain the marking information of the bottom surface image.

[0041] In the technical solution of this application embodiment, different filling methods are used to fill the initial degummed area for different wetting characteristics, so that the identification of the target degummed area is more accurate.

[0042] In some embodiments, the debonding detection result includes the debonding ratio. The debonding detection result of the battery pack is determined based on the marking information within the bottom image, including:

[0043] For each degummed battery cell, the degumming ratio of the degummed battery cell is calculated based on the marking information in the bottom image and the bottom area of ​​the degummed battery cell.

[0044] The degumming ratio of the battery pack is calculated based on the degumming ratio of multiple degummed cells.

[0045] In the technical solution of this application embodiment, the degumming ratio of the entire battery pack can be accurately determined based on the degumming ratio of a single degummed cell. Compared with manual visual inspection, this can reduce the influence of subjective factors, improve detection accuracy, and meet the quantitative analysis needs of degummed area in complex scenarios.

[0046] Secondly, this application also provides a battery pack delamination detection device for detecting whether delamination has occurred at the bottom of the battery cells in the battery pack. The device includes:

[0047] The position detection module is used to perform target detection on the initial image of the battery pack using a pre-trained position detection model to obtain the position information of each cell in the battery pack; wherein, in the initial image, each cell is arranged in its original position before the battery pack is disassembled and has been impregnated with a color developer;

[0048] The marker recognition module is used to perform region segmentation based on location information and the wetting color of the color developer, determine the bottom image of the debonded battery cell, and perform semantic analysis based on the wetting color of the color developer to identify the marker information in the bottom image.

[0049] The result determination module is used to determine the debonding detection result of the battery pack based on the marking information in the bottom image.

[0050] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method of any one of the first aspects.

[0051] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method of any one of the first aspects.

[0052] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the method of any one of the first aspects. Attached Figure Description

[0053] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the alternative embodiments below. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0054] Figure 1 This is a schematic diagram of the application environment of a battery pack delamination detection method according to an embodiment of this application;

[0055] Figure 2 This is a schematic flowchart of a battery pack debonding detection method according to an embodiment of this application;

[0056] Figure 3 This is a schematic flowchart of the position detection steps according to an embodiment of this application;

[0057] Figure 4 This is a schematic diagram of the coordinates of a corner point in the initial image of an embodiment of this application;

[0058] Figure 5 This is a flowchart illustrating the steps of determining the bottom surface image according to an embodiment of this application;

[0059] Figure 6 This is a flowchart illustrating the region segmentation step based on preliminary detection results according to an embodiment of this application;

[0060] Figure 7 This is a bottom image of a battery cell according to an embodiment of this application;

[0061] Figure 8 This is a flowchart illustrating the steps for determining marker information according to an embodiment of this application;

[0062] Figure 9a This is an image of the filling process according to an embodiment of this application;

[0063] Figure 9bThis is one of the schematic diagrams of the wetting features according to an embodiment of this application;

[0064] Figure 9c This is a schematic diagram of the filled impregnation area according to an embodiment of this application;

[0065] Figure 9d This is a second schematic diagram of the wetting feature according to an embodiment of this application;

[0066] Figure 9e This is a third schematic diagram of the wetting feature according to an embodiment of this application;

[0067] Figure 10 This is a flowchart illustrating the steps for determining the degumming test result according to an embodiment of this application;

[0068] Figure 11 This is a structural block diagram of a battery pack debonding detection device according to an embodiment of this application;

[0069] Figure 12 This is an internal structural diagram of a computer device according to an embodiment of this application. Detailed Implementation

[0070] The embodiments of the technical solution of this application will now be described in detail with reference to the accompanying drawings. These embodiments are only used to more clearly illustrate the technical solution of this application and are therefore merely examples, and should not be used to limit the scope of protection of this application.

[0071] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0072] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.

[0073] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0074] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0075] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).

[0076] In the description of the embodiments of this application, unless otherwise expressly specified and limited, technical terms such as "installation," "connection," "joining," and "fixing" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. For those skilled in the art, the specific meaning of the above terms in the embodiments of this application can be understood according to the specific circumstances.

[0077] With the widespread application of lithium-ion batteries in new energy vehicles, energy storage systems, and other fields, the reliability and safety of battery packs have become a key concern. In practical use, battery packs may be subjected to vibration and impact, causing the cells to delaminate against the base plate, thus affecting the performance and lifespan of the battery pack.

[0078] Currently, the method for detecting debonding involves using red ink as a colorimetric agent. After disassembling the battery pack, the area of ​​debonding is determined manually by visual inspection. However, manual inspection is time-consuming and cannot meet the needs of large-scale testing, resulting in low testing efficiency. Furthermore, manual inspection is greatly affected by subjective factors, making it difficult to guarantee the accuracy of the test results, leading to insufficient testing precision.

[0079] To address the aforementioned problems, this application provides a method for detecting battery pack delamination. This method utilizes a pre-trained position detection model to perform target detection on an initial image of the battery pack, obtaining the position information of each cell within the battery pack. In the initial image, each cell is arranged in its original position before disassembly and is impregnated with a color developer. Based on the position information and the impregnation color of the color developer, region segmentation is performed to determine the bottom image of the delamination cell. Semantic analysis is then performed based on the impregnation color of the color developer to identify marker information within the bottom image. Based on the marker information within the bottom image, the delamination detection result of the battery pack is determined. In the technical solution of this application, automatic image-based delamination detection can meet the needs of large-scale detection and improve detection efficiency. Furthermore, this solution first accurately locates the disassembled cells, reducing inaccuracies caused by factors such as shooting posture and irregular cell arrangement. Extracting the bottom image based on the location further improves the detection accuracy of the delamination area, thereby improving the detection precision of battery pack delamination.

[0080] The battery pack delamination detection method provided in this application embodiment can be applied to, for example... Figure 1 The application environment shown includes a terminal 101 and an image acquisition device 102. After the battery pack is impregnated with a color developer, the battery pack is disassembled, and the disassembled cells are arranged in their original positions. Then, the image acquisition device 102 acquires an initial image. The terminal 101 communicates with the image acquisition device 102 to obtain the initial image. The terminal 101 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, and standalone physical servers. The image acquisition device can be, but is not limited to, various cameras and webcams.

[0081] According to some embodiments of this application, refer to Figure 2 A method for detecting battery pack delamination is provided, which can be applied to... Figure 1 Taking the terminal in the example, the following steps can be included:

[0082] Step 201: Use a pre-trained position detection model to perform target detection on the initial image of the battery pack to obtain the position information of each cell in the battery pack.

[0083] In the initial image, each cell is arranged in its original position before the battery pack was disassembled and has been impregnated with a color developer.

[0084] After impregnating the battery pack with a color developer, the battery pack was disassembled, and the disassembled cells were arranged in their original positions before disassembly. Then, an image acquisition device was used to acquire initial images of the disassembled cells.

[0085] The terminal acquires an initial image from the image acquisition device. Then, the terminal inputs the initial image into a pre-trained position detection model. The position detection model detects the position of each battery cell in the initial image and then outputs the position information of each battery cell in the battery pack.

[0086] In some embodiments, the position detection model considers the influence of factors such as shooting posture and irregular cell arrangement during the detection process, corrects the position of each cell, and outputs the corrected position information of each cell.

[0087] In some embodiments, the location detection model can employ a lightweight object detection model. The process of training the location detection model may include: acquiring a training set, which includes multiple training images and annotations for each training image. The training images are images of battery cells arranged in their original positions after the battery pack has been immersed and disassembled. The annotations are the corner coordinates of each battery cell in the training images. The training images are input into the object detection model to obtain the location detection results output by the object detection model. Loss is calculated based on the location detection results and annotations, and the parameters of the object detection model are adjusted based on the calculated loss value. Training ends when the calculated loss value meets the convergence condition, and the object detection model at the end of training is determined as the location detection model.

[0088] Step 202: Based on the location information and the wetting color of the color developer, the region is segmented to determine the bottom image of the debonded battery cell, and semantic analysis is performed based on the wetting color of the color developer to identify the marking information in the bottom image.

[0089] The marking information is used to mark the debonded areas of the debonded battery cells.

[0090] Based on the location information of each battery cell in the battery pack, the terminal can determine the location of each cell in the initial image. If the area where the cell is located detects the color of the developer, the cell is identified as a debonded cell. For example, if the developer is red ink, and red is detected in the area where the cell is located, the cell is identified as a debonded cell. Then, the initial image is segmented based on the location of the debonded cell to obtain the bottom image of the debonded cell.

[0091] The required pixel value range for detection is determined in advance based on the wetting color of the color developer. Semantic analysis is performed on the bottom image of each debonded battery cell to select target pixels in the bottom image whose pixel values ​​fall within the above-mentioned pixel value range; the marking information in the bottom image is obtained based on the region where the target pixels are located.

[0092] Step 203: Determine the debonding detection result of the battery pack based on the marking information in the bottom image.

[0093] The battery pack debonding test results include the debonding area of ​​the battery pack.

[0094] The marking information in the bottom image indicates the target degummed area of ​​each degummed cell. Based on the target degummed area of ​​each degummed cell, the degummed area of ​​each degummed cell is calculated. The sum of the degummed areas of all degummed cells is calculated to obtain the degummed area of ​​the battery pack.

[0095] In some embodiments, calculating the debonded area of ​​each debonded battery cell based on the target debonded area of ​​each debonded battery cell may include: determining the correspondence between pixel size and actual size based on the device parameters of the image acquisition device and the distance between the image acquisition device and the battery cell; counting the number of pixels in the target debonded area; and calculating the debonded area based on the correspondence between pixel size and actual size, pixel size, and pixel number.

[0096] In the above embodiments, a pre-trained position detection model is used to perform target detection on the initial image of the battery pack to obtain the position information of each cell in the battery pack; based on the position information and the wetting color of the color developer, region segmentation is performed to determine the bottom image of the debonded cell, and semantic analysis is performed based on the wetting color of the color developer to identify the marking information in the bottom image; based on the marking information in the bottom image, the debonding detection result of the battery pack is determined. In the technical solution of this application embodiment, using images for automatic debonding detection can meet the needs of large-scale detection and improve detection efficiency; furthermore, this solution first accurately locates the disassembled cells, which can reduce the detection inaccuracy caused by factors such as shooting posture and irregular cell arrangement; extracting the bottom image based on the location can further improve the detection accuracy of the debonded area, thereby improving the detection accuracy of battery pack debonding.

[0097] According to some embodiments of this application, refer to Figure 3 In the above embodiment, "using a pre-trained position detection model to perform target detection on the initial image of the battery pack to obtain the position information of each cell in the battery pack" may include the following steps:

[0098] Step 301: Input the initial image into the position detection model for target detection to obtain the corner coordinates of multiple battery cells output by the position detection model.

[0099] The terminal inputs the initial image into the position detection model. The position detection model detects the position of each battery cell in the initial image and outputs the corner coordinates of multiple battery cells, referring to... Figure 4 .

[0100] Step 302: Map the corner coordinates of multiple cells and determine the position information of each cell in the battery pack based on the mapping results.

[0101] Due to factors such as shooting posture and irregular arrangement of disassembled battery cells, the shape of the battery cells in the initial image may be distorted, and the position detection model may also miss the corner coordinates of some cells. In this case, an N*M two-dimensional grid can be set according to the battery pack structure. After obtaining the corner coordinates of multiple cells, the initially located cells are mapped onto this two-dimensional grid. Under normal circumstances, the cells in the two-dimensional grid are arranged closely and with regular spacing. If there are empty spaces in the two-dimensional grid, it is determined that the position detection model has missed some cells. In this case, the corner coordinates of the missed cells are calculated based on the corner coordinates of adjacent cells. Finally, the position information of each cell in the battery pack is determined based on the corner coordinates of multiple cells output by the position detection model and the corner coordinates of the mapped and completed cells. This position information can be represented by a matrix.

[0102] In the above embodiments, the initial image is input into the position detection model for target detection, and the corner coordinates of multiple battery cells output by the position detection model are obtained. The corner coordinates of the multiple battery cells are then mapped, and the position information of each battery cell in the battery pack is determined based on the mapping result. In the technical solution of this application embodiment, the use of a position detection model and two-dimensional mesh mapping processing can reduce the influence of shooting posture and battery cell arrangement, reduce the dependence of shooting on tooling fixtures, and improve the robustness of the overall detection process.

[0103] According to some embodiments of this application, refer to Figure 5 In the above embodiments, "region segmentation based on location information and the wetting color of the color developer to determine the bottom image of the debonded battery cell" may include the following steps:

[0104] Step 401: Determine the location of each cell in the initial image based on the location information of each cell in the battery pack.

[0105] Based on the corner coordinates of each cell in the battery pack, the bounding box of each cell is located in the initial image; the region of each cell is then divided based on the bounding box.

[0106] Step 402: Detect the area of ​​each battery cell based on the wetting color of the color developer to obtain the preliminary detection results for each battery cell.

[0107] The preliminary test results are used to characterize whether the battery cell has delamination.

[0108] The immersion color of the color developer is characterized using HSV. HSV is a color space based on human color perception, short for Hue, Saturation, and Value. It decomposes color into three independent dimensions. Taking red ink as an example, the range of red is represented by HSV as follows: Hue (H): 0°-15° or 345°-360° (a fixed range of red on the color wheel), Saturation (S): 50%-100% (the higher the value, the purer the red; the lower the value, the closer it is to gray), Value (V): 50%-100% (the higher the value, the brighter the red; the lower the value, the closer it is to black).

[0109] For each cell, if the HSV of the area where the cell is located is within the above range, it is determined that red ink has seeped between the cell and the base plate, and the cell has delamination. If the HSV of the area where the cell is located is not within the above range, it is determined that the cell has not delamination.

[0110] Step 403: Based on the preliminary detection results, the initial image is segmented to determine the bottom image of the debonded battery cell.

[0111] The initial image is segmented based on the corner coordinates of each cell to obtain bottom images of multiple cells. Based on the initial detection results, the bottom images of multiple cells are filtered to identify the bottom images of cells with delamination, thus obtaining the bottom images of delamination cells.

[0112] In the above embodiments, based on the location information of each cell in the battery pack, the location of each cell is determined in the initial image; the location of each cell is detected based on the wetting color of the color developer to obtain the preliminary detection result corresponding to each cell; the initial image is segmented based on the preliminary detection result to determine the bottom image of the debonded cell. In the technical solution of this application embodiment, the initial screening based on the wetting color of the color developer can identify debonded cells, facilitating subsequent detection of only debonded cells and not non-debonded cells, thereby improving detection efficiency.

[0113] According to some embodiments of this application, refer to Figure 6 In the above embodiment, "performing region segmentation of the initial image based on preliminary detection results to determine the bottom image of the debonded battery cell" includes:

[0114] Step 501: Based on the preliminary detection results, the initial image is cropped to obtain a regional image of the area where the debonded battery cell is located.

[0115] Based on the preliminary detection results, the location of the debonded battery cell was determined. This area was then cropped out from the initial image, resulting in a regional image of the debonded battery cell. After cropping, the regional image can be normalized, for example, to a fixed resolution of 256*256.

[0116] Step 502: Use a pre-trained bottom surface recognition model to recognize and process the region image to obtain the bottom surface image of the debonded battery cell.

[0117] Since the region image includes the bottom surface of the battery cell and the area near the bottom surface, the region image is input into a pre-trained bottom surface recognition model, which is then used to identify the bottom surface image of the debonded battery cell from the region image.

[0118] In some embodiments, using a pre-trained bottom surface recognition model to process a region image to obtain a bottom surface image of the debonded battery cell may include: inputting the region image into the bottom surface recognition model for recognition processing to obtain the bottom surface position of the debonded battery cell; and cropping the region image according to the bottom surface position of the debonded battery cell to obtain the bottom surface image of the debonded battery cell. (Refer to...) Figure 7 The left, middle, and right images are regional images, with a bottom image superimposed on the regional images, and the cropped bottom image.

[0119] In some embodiments, the bottom surface recognition model can employ the ResUnet segmentation model. The process of training the bottom surface recognition model may include: obtaining a training set, which includes multiple training images and annotations for each training image. The training images are bottom surface images of a single battery cell, and the annotations are the bottom surface positions in each bottom surface image of the training images; inputting the training images into the segmentation model to obtain the bottom surface positions output by the segmentation model; calculating the loss based on the bottom surface positions and annotations, and adjusting the parameters of the segmentation model based on the calculated loss value. Training ends when the calculated loss value meets the convergence condition, and the segmentation model at the end of training is determined as the bottom surface recognition model.

[0120] In the above embodiments, the initial image is cropped based on the preliminary detection results to obtain a region image of the area where the debonded battery cell is located; the region image is then processed using a pre-trained bottom surface recognition model to obtain a bottom surface image of the debonded battery cell. In the technical solution of this application embodiment, using the cropped region image for bottom surface recognition reduces the computational burden and training difficulty of the bottom surface recognition model; using the bottom surface recognition model reduces the impact of image deformation on recognition accuracy, thereby enhancing the accuracy of bottom surface image extraction. This is particularly suitable for soft-pack batteries or irregularly shaped batteries, effectively separating the bottom surface image even if the battery cell has an irregular shape or slight adhesion.

[0121] According to some embodiments of this application, refer to Figure 8 In the above embodiments, "semantic analysis based on the wetting color of the developer to identify the marking information in the bottom image" may include the following steps:

[0122] Step 601: For each debonded battery cell, perform color conversion processing on the bottom image to obtain a converted image.

[0123] For each debonded battery cell, the bottom image of the debonded battery cell is converted from the RGB color space to the HSV color space to obtain the converted image.

[0124] Step 602: Perform semantic analysis on the converted image based on the wetting color of the developer to identify the marking information in the bottom image.

[0125] The marking information indicates the target debonding area of ​​the debonded battery cell.

[0126] A pre-trained color semantic analysis model is used to perform semantic analysis on the converted image to determine the target degummed area in the bottom image.

[0127] In some embodiments, semantic analysis of the converted image based on the wetting color of the developer to identify the marking information in the bottom image may include: performing identification processing on the converted image based on the wetting color of the developer to obtain the initial degummed area in the converted image; determining the wetting characteristics of the initial degummed area, and performing filling processing based on the wetting characteristics to obtain the marking information of the bottom image.

[0128] Taking red ink as the developer as an example, the range of red is represented by HSV: Hue (H): 0°-15° or 345°-360°, Saturation (S): 50%-100%, Lightness (V): 50%-100%. For each pixel in the converted image, if the HSV of that pixel is within the above range, then that pixel is determined to be stained by the developer. This process is repeated to determine whether each pixel is stained by the developer. Then, morphological operations, such as dilation, erosion, and connectivity adjustments, are performed on the pixels stained by the developer in the converted image to obtain the initial degummed region.

[0129] The developer penetrates from the edge of the bottom surface of the battery cell towards its center. Normally, the debonded areas on the bottom surface of the cell are stained by the developer, while the non-debonded areas remain unstained, resulting in a continuous penetration path. However, due to factors such as shooting posture and the degree of developer penetration, the penetration path in the bottom image may be discontinuous or incomplete.

[0130] The terminal first determines the wetting characteristics of the initial debonded area based on its shape. If abnormal wetting characteristics exist, the initial debonded area is filled to obtain the marking information of the bottom image, making the identification of the debonded area more accurate. Figure 9a .

[0131] Depending on the different wetting characteristics, the filling treatment may include the following:

[0132] 1) When the wetting feature is that the wetting area does not include the edge of the battery cell and a closed loop is formed at the center of the bottom surface, the outer ring at the center of the bottom surface is filled to obtain the marking information of the bottom surface image.

[0133] If the adhesive has detached from the edge of the cell's bottom surface but not from the center, the wetting area should include the bottom edge and form a closed loop at the center. However, referring to... Figure 9b In the converted image, the infiltration region forms a closed loop at the center of the bottom surface, but does not include the edge of the battery cell. To address this unusual infiltration feature, the outer ring at the center of the bottom surface is filled with the infiltration region, referring to... Figure 9c Then, based on all the infiltrated areas, the marking information of the bottom surface image is obtained.

[0134] 2) When the wetting feature is that the wetting area includes the edge of the battery cell but there are gaps, the area between the center of the bottom surface and the edge of the battery cell is filled to obtain the marking information of the bottom surface image.

[0135] Reference Figure 9d The image shows a wetted region encompassing the cell edge; however, a gap exists between the wetted region and the center of the bottom surface. To address this unusual wetted feature, the area between the center of the bottom surface and the cell edge is filled with the wetted region, as shown in the reference image. Figure 9c Then, based on all the infiltrated areas, the marking information of the bottom surface image is obtained.

[0136] 3) When the wetting feature is that the wetting area does not include the edge of the battery cell and a closed loop is not formed at the center of the bottom surface, the unclosed position at the center of the bottom surface is closed, and the outer loop of the closed center of the bottom surface is filled to obtain the marking information of the bottom surface image.

[0137] Reference Figure 9e The infiltration region in the converted image does not include the cell edge, and no closed loop is formed at the center of the bottom surface. To address this abnormal infiltration characteristic, the center of the bottom surface is first closed, forming a closed loop of the infiltration region at the center. Then, the outer loop of the closed bottom surface center is filled with the infiltration region, referring to... Figure 9c Then, based on all the infiltrated areas, the marking information of the bottom surface image is obtained.

[0138] In the above embodiments, for each debonded battery cell, the bottom image is color-converted to obtain a converted image; semantic analysis is performed on the converted image based on the wetting color of the developer to identify the marking information within the bottom image. In the technical solution of this application embodiment, the debonded area is reconstructed, improving the completeness of identification of weak or localized impregnation and reducing the risk of missed detection.

[0139] According to some embodiments of this application, the degumming test results include the degumming ratio, as referred to... Figure 10 In the above embodiments, "determining the debonding detection result of the battery pack based on the marking information in the bottom image" may include the following steps:

[0140] Step 701: For each degummed battery cell, calculate the degumming ratio of the degummed battery cell based on the marking information in the bottom image and the bottom area of ​​the degummed battery cell.

[0141] For each debonded battery cell, the terminal can calculate the bottom area of ​​the debonded battery cell based on the number of pixels in the bottom image of the cell and the pre-established correspondence between pixel size and actual size. The debonded area is calculated by determining the number of pixels, pixel size, and correspondence between pixel size and actual size within the target debonded area based on the marking information in the bottom image; then, the ratio of the debonded area to the bottom area is calculated to obtain the debonding ratio of the debonded battery cell.

[0142] In some embodiments, the number of pixels in the target degummed area and the number of pixels in the bottom image can be determined first, and then the ratio of the two can be calculated to obtain the degummed ratio of the degummed battery cell.

[0143] Step 702: Calculate the degumming ratio of the battery pack based on the degumming ratio of multiple degummed cells.

[0144] After calculating the degumming ratio of each degummed cell, the largest degumming ratio can be determined as the degumming ratio of the battery pack, or the median of multiple degumming ratios can be determined as the degumming ratio of the battery pack; alternatively, the average of multiple degumming ratios can be calculated and the average value can be determined as the degumming ratio of the battery pack.

[0145] In the above embodiments, for each debonded cell, the debonding ratio of the debonded cell is calculated based on the marking information in the bottom image and the bottom area of ​​the debonded cell; the debonding ratio of the battery pack is calculated based on the debonding ratios of multiple debonded cells. In the technical solution of this application embodiment, the debonding ratio of the entire battery pack can be accurately determined based on the debonding ratio of a single debonded cell. Compared with manual visual inspection, this reduces the influence of subjective factors, improves detection accuracy, and meets the quantitative analysis needs of debonded area in complex scenarios.

[0146] According to some embodiments of this application, a method for detecting battery pack delamination is provided, which can be applied to... Figure 1 Taking the terminal in the example, the following steps can be included:

[0147] Step 1: Obtain the initial image and input it into the pre-trained position detection model to obtain the corner coordinates of multiple battery cells output by the position detection model.

[0148] Step 2: Map the corner coordinates of multiple cells and determine the position information of each cell in the battery pack based on the mapping results.

[0149] Step 3: Based on the location information of each cell in the battery pack, determine the location of each cell in the initial image.

[0150] Step 4: Detect the area of ​​each cell based on the wetting color of the color developer to obtain the preliminary detection results for each cell; the preliminary detection results are used to characterize whether the cell has delamination.

[0151] Step 5: Based on the preliminary detection results, the initial image is cropped to obtain a regional image of the area where the debonded battery cell is located.

[0152] Step 6: Input the region image into the pre-trained bottom surface recognition model for recognition processing to obtain the bottom surface position of the debonded battery cell.

[0153] Step 7: Crop the area image according to the bottom position of the debonded battery cell to obtain the bottom image of the debonded battery cell.

[0154] Step 8: For each debonded battery cell, perform color conversion processing on the bottom image to obtain the converted image.

[0155] Step 9: Identify and process the converted image based on the wetting color of the developer to obtain the initial degummed area in the converted image.

[0156] Step 10: Determine the wetting characteristics of the initial degummed area, and perform filling processing based on the wetting characteristics to obtain the marking information in the bottom image of the degummed battery cell.

[0157] Specifically, when the wetting feature is that the wetting area does not include the edge of the battery cell and a closed loop is formed at the center of the bottom surface, the outer ring at the center of the bottom surface is filled to obtain the marking information of the bottom surface image; when the wetting feature is that the wetting area includes the edge of the battery cell but there are gaps, the area between the center of the bottom surface and the edge of the battery cell is filled to obtain the marking information of the bottom surface image; when the wetting feature is that the wetting area does not include the edge of the battery cell and a closed loop is not formed at the center of the bottom surface, the unclosed position at the center of the bottom surface is closed, and the outer ring at the closed center of the bottom surface is filled to obtain the marking information of the bottom surface image.

[0158] Step 11: For each degummed battery cell, calculate the degumming ratio of the degummed battery cell based on the marking information in the bottom image and the bottom area of ​​the degummed battery cell.

[0159] Step 12: Calculate the degumming ratio of the battery pack based on the degumming ratio of multiple degummed cells.

[0160] The technical solution of this application combines target detection, segmentation algorithms, and color semantic analysis to construct a visual inspection process for battery cell delamination that adapts to complex shooting conditions. This process can accurately identify the delamination area of ​​a single battery cell even when cells are misaligned, have inconsistent shapes, or have missed red ink penetration, thus accurately outputting the delamination area of ​​the entire battery pack. Furthermore, the automatic delamination detection can meet the needs of large-scale inspection and improve inspection efficiency.

[0161] It should be understood that although the steps in the flowchart above are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowchart above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps.

[0162] Based on the same inventive concept, this application also provides a battery pack debonding detection device for implementing the battery pack debonding detection method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more embodiments of the battery pack debonding detection device provided below can be found in the limitations of the battery pack debonding detection method described above, and will not be repeated here.

[0163] According to some embodiments of this application, refer to Figure 11 A battery pack delamination detection device is provided, the device comprising:

[0164] The position detection module 801 is used to perform target detection on the initial image of the battery pack using a pre-trained position detection model to obtain the position information of each cell in the battery pack; wherein, in the initial image, each cell is arranged in its original position before the battery pack is disassembled and has been impregnated with a color developer.

[0165] The marker recognition module 802 is used to perform region segmentation based on location information and the wetting color of the color developer, determine the bottom image of the debonded battery cell, and perform semantic analysis based on the wetting color of the color developer to identify the marker information in the bottom image.

[0166] The result determination module 803 is used to determine the debonding detection result of the battery pack based on the marking information in the bottom image.

[0167] In some embodiments, the position detection module 801 is specifically used to input the initial image into the position detection model for target detection, obtain the corner coordinates of multiple battery cells output by the position detection model, perform mapping processing on the corner coordinates of multiple battery cells, and determine the position information of each battery cell in the battery pack based on the mapping result.

[0168] In some embodiments, the marker recognition module 802 is specifically used to determine the location of each cell in the initial image based on the location information of each cell in the battery pack; to detect the location of each cell based on the wetting color of the color developer, and to obtain the preliminary detection result corresponding to each cell; wherein the preliminary detection result is used to characterize whether the cell has delamination; and to perform region segmentation on the initial image based on the preliminary detection result to determine the bottom image of the delamination cell.

[0169] In some embodiments, the marker recognition module 802 is specifically used to crop the initial image based on the preliminary detection results to obtain a region image of the area where the debonded battery cell is located; and to use a pre-trained bottom surface recognition model to recognize the region image to obtain a bottom surface image of the debonded battery cell.

[0170] In some embodiments, the marker recognition module 802 is specifically used to input the region image into a pre-trained bottom surface recognition model for recognition processing to obtain the bottom surface position of the debonded battery cell; and to crop the region image according to the bottom surface position of the debonded battery cell to obtain the bottom surface image of the debonded battery cell.

[0171] In some embodiments, the marker recognition module 802 is specifically used to perform color conversion processing on the bottom image of each debonded battery cell to obtain a converted image; and to perform semantic analysis on the converted image based on the wetting color of the color developer to identify the marker information in the bottom image.

[0172] In some embodiments, the marker recognition module 802 is specifically used to recognize the converted image according to the wetting color of the developer to obtain the initial degummed area in the converted image; determine the wetting characteristics of the initial degummed area, and perform filling processing according to the wetting characteristics to obtain the marker information of the bottom image.

[0173] In some embodiments, the marker recognition module 802 is specifically used to fill the outer ring at the center of the bottom surface when the immersion feature is that the immersion area does not contain the edge of the battery cell and a closed loop is formed at the center of the bottom surface, so as to obtain the marker information of the bottom surface image.

[0174] In some embodiments, the marker recognition module 802 is specifically used to fill the area between the bottom center and the edge of the battery cell when the immersion feature is that the immersion area includes the edge of the battery cell but there are gaps, so as to obtain the marker information of the bottom image.

[0175] In some embodiments, the marker recognition module 802 is specifically used to close the unclosed position at the center of the bottom surface when the immersion feature is that the immersion area does not contain the edge of the battery cell and a closed loop is not formed at the center of the bottom surface, and to fill the outer ring of the closed center of the bottom surface to obtain the marker information of the bottom surface image.

[0176] In some embodiments, the degumming detection result includes the degumming ratio. The result determination module 803 is specifically used to calculate the degumming ratio of each degummed cell based on the marking information in the bottom image and the bottom area of ​​the degummed cell; and to calculate the degumming ratio of the battery pack based on the degumming ratio of multiple degummed cells.

[0177] Each module in the aforementioned battery pack delamination detection device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of the electronic device in hardware form or independent of it, or stored in the memory of the electronic device in software form, so that the processor can call and execute the corresponding operations of each module.

[0178] According to some embodiments of this application, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 12As shown, the computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When executed by the processor, the computer program implements a method for detecting battery pack delamination. The display unit is used to form a visually visible image and can be a display screen, projection device, or virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0179] Those skilled in the art will understand that Figure 12 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0180] According to some embodiments of this application, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory including instructions that can be executed by a processor of an electronic device to perform the above-described method. For example, the non-transitory computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc.

[0181] According to some embodiments of this application, a computer program product is also provided, which, when executed by a processor, can implement the above-described methods. The computer program product includes one or more computer instructions. When these computer instructions are loaded and executed on a computer, some or all of the above-described methods can be implemented, wholly or partially, according to the processes or functions described in the embodiments of this application.

[0182] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0183] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0184] The embodiments described above merely illustrate several implementation methods of this application to facilitate a detailed understanding of the technical solutions of this application, but should not be construed as limiting the scope of protection of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the scope of protection of this application. It should be understood that technical solutions obtained by those skilled in the art based on the technical solutions provided in this application through logical analysis, reasoning, or limited experimentation are all within the scope of protection of the appended claims. Therefore, the scope of protection of this patent application should be determined by the content of the appended claims, and the specification and drawings can be used to interpret the content of the claims.

Claims

1. A method for detecting delamination in a battery pack, used to detect whether delamination has occurred at the bottom of the battery cells in the battery pack, characterized in that, The method includes: The initial image of the battery pack is subjected to target detection using a pre-trained position detection model to obtain the position information of each cell in the battery pack; wherein, in the initial image, each cell is arranged in its original position before the battery pack is disassembled and has been impregnated with a color developer; Based on the location information and the wetting color of the color developer, region segmentation is performed to determine the bottom image of the debonded battery cell, and semantic analysis is performed based on the wetting color of the color developer to identify the marking information in the bottom image. Based on the marking information in the bottom image, the debonding detection result of the battery pack is determined; The step of using a pre-trained position detection model to perform target detection on the initial image of the battery pack to obtain the position information of each cell in the battery pack includes: The initial image is input into the position detection model for target detection, and the corner coordinates of multiple battery cells are obtained from the position detection model. The corner coordinates of multiple battery cells are mapped, and the position information of each battery cell in the battery pack is determined based on the mapping result.

2. The method according to claim 1, characterized in that, The step of segmenting regions based on the location information and the wetting color of the color developer to determine the bottom image of the debonded battery cell includes: Based on the location information of each cell in the battery pack, the location of each cell is determined in the initial image; The area where each of the battery cells is located is detected based on the wetting color of the color developer to obtain preliminary detection results for each battery cell; wherein, the preliminary detection results are used to characterize whether the battery cell has delamination. Based on the preliminary detection results, the initial image is segmented to determine the bottom image of the debonded battery cell.

3. The method according to claim 2, characterized in that, The step of segmenting the initial image based on the preliminary detection results to determine the bottom image of the debonded battery cell includes: Based on the preliminary detection results, the initial image is cropped to obtain a region image of the area where the debonded battery cell is located. The bottom surface image of the debonded battery cell is obtained by using a pre-trained bottom surface recognition model to recognize and process the image of the region.

4. The method according to claim 3, characterized in that, The step of using a pre-trained bottom surface recognition model to recognize and process the region image to obtain the bottom surface image of the debonded battery cell includes: The image of the region is input into the bottom surface recognition model for recognition processing to obtain the bottom surface position of the debonded battery cell; The image of the region is cropped based on the position of the bottom surface of the debonded battery cell to obtain the bottom surface image of the debonded battery cell.

5. The method according to claim 1, characterized in that, The semantic analysis based on the wetting color of the color developer to identify the marker information within the bottom image includes: For each of the debonded battery cells, the bottom image is color-converted to obtain a converted image; Based on the wetting color of the developer, semantic analysis is performed on the converted image to identify the marking information within the bottom image.

6. The method according to claim 5, characterized in that, The step of performing semantic analysis on the converted image based on the wetting color of the developer to identify the marker information within the bottom image includes: The conversion image is identified and processed based on the wetting color of the color developer to obtain the initial degummed area in the conversion image; The wetting characteristics of the initial degummed area are determined, and the marking information of the bottom image is obtained by filling the wetting characteristics.

7. The method according to claim 6, characterized in that, The step of obtaining the marking information of the bottom image by filling based on the infiltration features includes: When the wetting feature is that the wetting area does not include the edge of the battery cell and forms a closed loop at the center of the bottom surface, the outer ring at the center of the bottom surface is filled to obtain the marking information of the bottom surface image.

8. The method according to claim 6, characterized in that, The step of obtaining the marking information of the bottom image by filling based on the infiltration features includes: When the wetting feature is that the wetting area includes the edge of the battery cell but has gaps, the area between the center of the bottom surface and the edge of the battery cell is filled to obtain the marking information of the bottom surface image.

9. The method according to claim 6, characterized in that, The step of obtaining the marking information of the bottom image by filling based on the infiltration features includes: When the wetting feature is that the wetting area does not include the edge of the battery cell and a closed loop is not formed at the center of the bottom surface, the unclosed position at the center of the bottom surface is closed, and the outer ring of the closed center of the bottom surface is filled to obtain the marking information of the bottom surface image.

10. The method according to any one of claims 1-9, characterized in that, The debonding detection result includes the debonding ratio. Determining the debonding detection result of the battery pack based on the marking information in the bottom image includes: For each of the degummed battery cells, the degumming ratio of the degummed battery cell is calculated based on the marking information in the bottom image and the bottom area of ​​the degummed battery cell; The degumming ratio of the battery pack is calculated based on the degumming ratio of the multiple degummed cells.

11. A battery pack delamination detection device, used to detect whether delamination has occurred at the bottom of the battery cells in the battery pack, characterized in that, The device includes: The position detection module is used to perform target detection on the initial image of the battery pack using a pre-trained position detection model to obtain the position information of each cell in the battery pack; wherein, in the initial image, each cell is arranged in its original position before the battery pack is disassembled and has been impregnated with a color developer; The marker recognition module is used to perform region segmentation based on the location information and the wetting color of the color developer to determine the bottom image of the debonded battery cell, and to perform semantic analysis based on the wetting color of the color developer to identify the marker information in the bottom image; The result determination module is used to determine the debonding detection result of the battery pack based on the marking information in the bottom image; Specifically, the position detection module is used to input the initial image into the position detection model for target detection, obtain the corner coordinates of multiple battery cells output by the position detection model, perform mapping processing on the corner coordinates of multiple battery cells, and determine the position information of each battery cell in the battery pack based on the mapping result.

12. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 10.

13. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 10.

14. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 10.

Images