Termination position inspection method and apparatus, and computer device and storage medium

Abstract

A termination position inspection method and apparatus, and a computer device and a storage medium. The inspection method comprises: in response to a termination position inspection instruction for a target battery cell, acquiring internal-structure images of the target battery cell from a plurality of different viewing angles (S201); and analyzing the internal-structure images, so as to determine a termination position inspection result of the target battery cell (S202). The accuracy of inspecting the termination position of a target battery cell can be improved.

Description

End position detection method, device, computer equipment and storage medium

[0001] Cross-references

[0002] This application refers to Chinese Patent Application No. 2024100738654, filed on January 18, 2024, entitled “End Position Detection Method, Device, Computer Equipment and Storage Medium”, which is incorporated into this application in its entirety by reference. Technical Field

[0003] The present application relates to the technical field of battery cell detection, and in particular to a method, apparatus, computer equipment, and storage medium for detecting a tail position. Background Art

[0004] With the development of electronic devices, batteries are used more and more widely.

[0005] Take wound batteries, for example. Due to their high energy density and lightweight nature, they have become the preferred power source for many electronic devices. However, during the production process, incorrect cell termination can lead to internal short circuits or damage. Therefore, defect detection at the cell termination point has become a pressing issue.

[0006] Summary of the Invention

[0007] Based on this, it is necessary to provide a tail position detection method, device, computer equipment and storage medium to address the above technical problems, which can detect defects in the tail position of the battery cell.

[0008] In a first aspect, an embodiment of the present application provides a method for detecting a tail position, comprising:

[0009] In response to a tail position detection instruction of a target battery cell, acquiring internal structure images of the target battery cell at multiple different viewing angles;

[0010] Analyze each internal structure image to determine the end position detection result of the target battery cell.

[0011] In the tail position detection method provided in an embodiment of the present application, in response to a tail position detection instruction of a target battery cell, internal structure images of the target battery cell at multiple different viewing angles are obtained, and then each internal structure image is analyzed to determine the tail position detection result of the target battery cell. In this method, the internal structure image can reflect the density distribution and composition inside the target battery cell. Therefore, the internal structure image of the target battery cell can be used to detect whether the tail position of the electrode inside the target battery cell is correct; and the tail position inside the target battery cell is detected by the internal structure image without destroying the integrity of the target battery cell, thereby achieving non-destructive detection of the tail position of the target battery cell; in addition, by determining the tail position detection result through the internal structure images of the target battery cell at multiple different viewing angles, it is possible to reduce the error that may exist in a single view, reduce the possibility of misjudgment, and improve the accuracy of the tail position detection result.

[0012] In one embodiment, the internal structure image includes an overall internal structure image of the target battery cell, and obtaining the internal structure images of the target battery cell from multiple different perspectives includes:

[0013] An overall internal structure image captured by a first image capture device is acquired, where the capture viewing angle of the first image capture device covers the entire surface of the target battery cell.

[0014] In the end position detection method provided in the embodiments of the present application, a first image acquisition device captures an overall internal structure image, and the first image acquisition device's capture angle of view covers the entire surface of the target battery cell. In this method, the first image acquisition device's capture angle of view covers the entire surface of the target battery cell, so that the captured overall internal structure image can represent all structural information within the target battery cell, facilitating a comprehensive analysis of the target battery cell's internal structure, thereby improving the accuracy of the end position detection results.

[0015] In one embodiment, the internal structure image includes an internal structure image of a corner of the target battery cell, and obtaining the internal structure images of the target battery cell from multiple different viewing angles includes:

[0016] Acquire a first internal structure image acquired by a second image acquisition device and a second internal structure image acquired by a third image acquisition device; the second image acquisition device has an acquisition angle covering an area where a first corner and a second corner of the target battery cell are located; the third image acquisition device has an acquisition angle covering an area where a third corner and a fourth corner of the target battery cell are located;

[0017] The first internal structure image and the second internal structure image are merged to obtain a corner internal structure image.

[0018] In the end position detection method provided in the embodiment of the present application, a first internal structure image captured by a second image acquisition device and a second internal structure image captured by a third image acquisition device are obtained, and the first internal structure image and the second internal structure image are merged to obtain a corner internal structure image; wherein, the capture angle of the second image acquisition device covers the area where the first corner and the second corner are located on the target battery cell; and the capture angle of the third image acquisition device covers the area where the third corner and the fourth corner are located on the target battery cell. In this method, since the end position of the target battery cell is detected in order to detect whether the end position of the battery cell is at the corner of the target battery cell, the end position of the corner internal structure image is detected in a targeted manner by capturing an image of the local corner area of ​​the target battery cell, thereby improving the accuracy and efficiency of the end position detection.

[0019] In one embodiment, each internal structure image includes an overall internal structure image and a corner internal structure image of the target battery cell; analyzing each internal structure image to determine a tail position detection result of the target battery cell includes:

[0020] Performing image enhancement processing on the overall internal structure image and the corner internal structure image respectively to obtain an enhanced overall internal structure image and an enhanced corner internal structure image;

[0021] The tail position detection result of the target battery cell is determined based on the overall internal structure image, the enhanced overall internal structure image and the enhanced corner internal structure image.

[0022] In the method for detecting the end position provided in an embodiment of the present application, image enhancement processing is performed on the overall internal structure image and the corner internal structure image, respectively, to obtain an enhanced overall internal structure image and an enhanced corner internal structure image. The end position detection result of the target battery cell is determined based on the overall internal structure image, the enhanced overall internal structure image, and the enhanced corner internal structure image. In this method, image enhancement processing is performed on the overall internal structure image and the corner internal structure image, which can highlight important features in the overall internal structure image and the corner internal structure image. Therefore, by detecting the end position using the overall internal structure image and the enhanced overall internal structure image and the enhanced corner internal structure image after enhancement processing, the accuracy of the end position detection result can be improved. Furthermore, by performing comprehensive detection of the end position using internal structure images of multiple dimensions, the accuracy of end position detection is further improved.

[0023] In one embodiment, performing image enhancement processing on the overall internal structure image and the corner internal structure image to obtain an enhanced overall internal structure image and an enhanced corner internal structure image respectively includes:

[0024] Determine the edge pixel points of the target image according to the grayscale value of each pixel point in the target image; the target image is an overall internal structure image or a corner internal structure image;

[0025] Obtain an image of the region of interest from the target image according to the edge pixels of the target image and a preset outward expansion offset;

[0026] Image processing is performed on the region of interest image to obtain an enhanced internal structure image of the target image.

[0027] In the end position detection method provided in an embodiment of the present application, the edge pixels of the target image are determined based on the grayscale value of each pixel in the target image; the target image is an overall internal structure image or a corner internal structure image; and based on the edge pixels of the target image and a preset outward expansion offset, a region of interest image is obtained from the target image, and then image processing is performed on the region of interest image to obtain an enhanced internal structure image of the target image. In this method, the region of interest images in the overall internal structure image and the corner internal structure image are first extracted, and then image processing is performed on the region of interest images. This reduces the computational complexity of enhancing the overall internal structure image or the corner internal structure image and improves image processing efficiency. Furthermore, when extracting the region of interest image, the region of interest is determined based on the edge pixels and the outward expansion offset, making the extracted region of interest more accurate and effective.

[0028] In one embodiment, performing image processing on the region of interest image to obtain an enhanced internal structure image of the target image includes:

[0029] Perform logarithmic domain transformation on the image of the region of interest to obtain the internal structure image in the logarithmic domain;

[0030] Performing bilateral filtering on the logarithmic domain internal structure image to obtain a filtered internal structure image;

[0031] The filtered internal structure image is equalized to obtain an enhanced internal structure image of the target image.

[0032] In the end position detection method provided in an embodiment of the present application, the region of interest image is transformed into a logarithmic domain to obtain a logarithmic domain internal structure image, and the logarithmic domain internal structure image is subjected to bilateral filtering to obtain a filtered internal structure image, and then the filtered internal structure image is equalized to obtain an enhanced internal structure image of the target image. In this method, the logarithmic domain transformation can improve the detail sensitivity of the image, the bilateral filtering can effectively reduce the noise in the image and improve the image quality, and the equalization processing can enhance the contrast of the image, making the grayscale changes in the image more obvious. Therefore, by performing this series of processing on the region of interest image, a clearer and more accurate internal structure image can be obtained, providing more accurate data support for subsequent end position detection.

[0033] In one embodiment, determining a detection result of a tail position of a target battery cell based on the overall internal structure image, the enhanced overall internal structure image, and the enhanced corner internal structure image includes:

[0034] The overall internal structure image, enhanced overall internal structure image, and enhanced corner internal structure image are all input into the position detection model to obtain the tail position detection result of the target battery cell.

[0035] In the tail position detection method provided in the embodiments of the present application, the overall internal structure image, the enhanced overall internal structure image, and the enhanced corner internal structure image are all input into a position detection model to obtain the tail position detection result of the target battery cell. In this method, the tail position detection result of the target battery cell is directly determined using a pre-trained position detection model, improving the efficiency and accuracy of tail position detection for the target battery cell.

[0036] In one embodiment, the position detection model includes a depthwise separable convolutional network and an output network; the overall internal structure image, the enhanced overall internal structure image, and the enhanced corner internal structure image are all input into the position detection model to obtain the tail position detection result of the target battery cell, including:

[0037] Through the depth-wise separable convolutional network, multi-scale feature extraction is performed on the overall internal structure image, enhanced overall internal structure image and enhanced corner internal structure image, respectively, to obtain the overall feature image, enhanced overall feature image and enhanced corner feature image at multiple different scales;

[0038] Each overall feature image, each enhanced overall feature image, and each enhanced corner feature image are input into the output network to obtain the tail position detection result.

[0039] In the end position detection method provided in the embodiments of the present application, a depthwise separable convolutional network is used to perform multi-scale feature extraction on the overall internal structure image, the enhanced overall internal structure image, and the enhanced corner internal structure image, respectively. This results in overall feature images, enhanced overall feature images, and enhanced corner feature images at multiple scales. Each of these overall feature images, enhanced overall feature images, and enhanced corner feature images is then input into an output network to obtain the end position detection result. This method determines the end position detection result of the target battery cell using feature images at different angles and multiple scales, thereby improving the accuracy of the end position detection result.

[0040] In one embodiment, determining a detection result of a tail position of a target battery cell based on the overall internal structure image, the enhanced overall internal structure image, and the enhanced corner internal structure image includes:

[0041] Determine a first probability value that the tail position of the target battery cell is normal based on the overall internal structure image and a preset standard overall internal structure image; determine a second probability value that the tail position of the target battery cell is normal based on the enhanced overall internal structure image and the preset standard enhanced overall internal structure image; determine a third probability value that the tail position of the target battery cell is normal based on the enhanced corner internal structure image and the preset standard enhanced corner internal structure image;

[0042] A detection result of the tail position of the target battery cell is determined according to the first probability value, the second probability value, and the third probability value.

[0043] In the tail position detection method provided in the embodiment of the present application, a first probability value that the tail position of the target battery cell is normal is determined based on the overall internal structure image and a preset standard overall internal structure image, and a second probability value that the tail position of the target battery cell is normal is determined based on the enhanced overall internal structure image and the preset standard enhanced overall internal structure image. Then, a third probability value that the tail position of the target battery cell is normal is determined based on the enhanced corner internal structure image and the preset standard enhanced corner internal structure image. Finally, based on the first probability value, the second probability value, and the third probability value, the tail position detection result of the target battery cell is determined. In this method, the overall internal structure image, the enhanced overall internal structure image, and the enhanced corner internal structure image are respectively compared with the corresponding standard internal structure image to determine multiple probability values ​​that the tail position of the target battery cell is normal. Then, the multiple probability values ​​are fused for decision making to determine the tail position detection result of the target battery cell, thereby improving the comprehensiveness of the tail position detection and thus improving the effectiveness and accuracy of the tail position detection result.

[0044] In one embodiment, determining a detection result of a target battery cell's end position according to the first probability value, the second probability value, and the third probability value includes:

[0045] Determining a comprehensive probability value based on the first probability value, the second probability value, and the third probability value;

[0046] When the comprehensive probability value is greater than the preset probability threshold, the tail position detection result is determined to be normal; when the comprehensive probability value is less than or equal to the probability threshold, the tail position detection result is determined to be abnormal.

[0047] In the end position detection method provided in an embodiment of the present application, a comprehensive probability value is determined based on a first probability value, a second probability value, and a third probability value. If the comprehensive probability value is greater than a preset probability threshold, the end position detection result is determined to be normal; if the comprehensive probability value is less than or equal to the probability threshold, the end position detection result is determined to be abnormal. In this method, a decision is made based on multiple probability values ​​indicating that the end position detection is normal, and the end position detection result of the target battery cell is determined based on the comprehensive probability value obtained from the decision and the preset probability threshold, thereby improving the accuracy of the end position detection.

[0048] In a second aspect, an embodiment of the present application further provides a tail position detection device, comprising:

[0049] An image acquisition module, configured to acquire internal structure images of the target battery cell from a plurality of different viewing angles in response to a tail position detection instruction of the target battery cell;

[0050] The result determination module is used to analyze each internal structure image and determine the end position detection result of the target battery cell.

[0051] In a third aspect, an embodiment of the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method provided in any embodiment of the first aspect when executing the computer program.

[0052] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method provided in any embodiment of the first aspect above.

[0053] In a fifth aspect, an embodiment of the present application further provides a computer program product, comprising a computer program, which, when executed by a processor, implements the steps of the method provided in any embodiment of the first aspect above.

[0054] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0056] FIG1 is a diagram showing the internal structure of a computer device according to an embodiment;

[0057] FIG2 is a schematic flow chart of a method for detecting a tail position in one embodiment;

[0058] FIG3 is a schematic diagram of the structure of an X-Ray device collecting an image of the entire internal structure in one embodiment;

[0059] FIG4 is a schematic flow chart of a method for detecting a tail position in another embodiment;

[0060] FIG5 is a schematic diagram of a structure in which an X-Ray device acquires a second internal structure image and a third internal structure image in one embodiment;

[0061] FIG6 is a schematic flow chart of a method for detecting a tail position in another embodiment;

[0062] FIG7 is a schematic flow chart of a method for detecting a tail position in another embodiment;

[0063] FIG8 is a schematic diagram of a structure for extracting an image of a region of interest in one embodiment;

[0064] FIG9 is a schematic flow chart of a method for detecting a tail position in another embodiment;

[0065] FIG10 is a schematic flow chart of a method for detecting a tail position in another embodiment;

[0066] FIG11 is a schematic diagram of the structure of a depthwise separable convolutional network according to one embodiment;

[0067] FIG12 is a schematic diagram of the structure of a YOLOv5 benchmark model according to one embodiment;

[0068] FIG13 is a schematic structural diagram of an initial position detection model in one embodiment;

[0069] FIG14 is a schematic flow chart of a method for detecting a tail position in another embodiment;

[0070] FIG15 is a schematic flow chart of a method for detecting a tail position in another embodiment;

[0071] FIG16 is a schematic diagram of X-Ray imaging of a corner area of ​​a target battery cell in one embodiment;

[0072] FIG17 is a schematic diagram of X-Ray imaging of a corner area of ​​a target battery cell in another embodiment;

[0073] FIG18 is a schematic diagram of an image of the overall internal structure of a battery cell in one embodiment;

[0074] FIG19 is a flow chart of a method for detecting a tail position in another embodiment;

[0075] FIG20 is a block diagram of a tail position detection device according to an embodiment;

[0076] FIG21 is a block diagram of a tail position detection device according to another embodiment;

[0077] FIG22 is a block diagram of a tail position detection device according to another embodiment;

[0078] FIG23 is a block diagram of a structure of a tail position detection device according to another embodiment;

[0079] FIG24 is a block diagram of a structure of a tail position detection device according to another embodiment;

[0080] FIG25 is a structural block diagram of a tail position detection device in another embodiment;

[0081] FIG26 is a block diagram of a tail position detection device according to another embodiment;

[0082] FIG27 is a block diagram of a tail position detection device according to another embodiment;

[0083] FIG28 is a block diagram of a structure of a tail position detection device according to another embodiment;

[0084] FIG29 is a structural block diagram of a tail position detection device in another embodiment.

[0085] The reference numerals in the specific implementation manner are as follows: 301 target battery cell; 302 large surface of the target battery cell; 303 first X-Ray device; 304 acquisition source; 305 flat panel detector; 501 area where the first corner is located; 502 area where the second corner is located; 503 area where the third corner is located; 504 area where the fourth corner is located; 505 second X-Ray device; 506 third X-Ray device; 801 overall internal structure image; 802 edge pixel points; 803 area of ​​interest image. DETAILED DESCRIPTION

[0086] In order to make the purpose, technical solutions and advantages of this application more clearly understood, the following embodiments of the technical solutions of this application are described in detail with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of this application and are therefore only examples and are not intended to limit the scope of protection of this application.

[0087] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned figure descriptions are intended to cover non-exclusive inclusions.

[0088] In the description of the embodiments of the present application, the technical terms "first" and "second" are only used to distinguish different objects, and cannot be understood as indicating or implying relative importance or implicitly indicating the number, specific order or primary and secondary relationship of the indicated technical features. In the description of the embodiments of the present application, "multiple" means more than two, unless otherwise clearly and specifically defined. Mentioning "embodiments" in this article means that the specific features, structures or characteristics described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0089] In the description of the embodiments of the present application, the term "plurality" refers to more than two (including two).

[0090] Taking lithium-ion batteries as an example, the assembly process typically involves assembling components such as the positive and negative electrode sheets, separators, tabs, and casing. This assembly process can be generally divided into winding and lamination, and assembly. Winding and lamination involves welding the positive and negative electrode sheets and separators with tabs welded to the current collectors to form a square or cylindrical battery cell structure with a positive electrode, separator, and negative electrode. Assembly involves assembling the battery cell, casing, cover, and insulating sheet.

[0091] The battery winding process typically involves arranging the positive and negative electrode sheets and separators in the order of positive electrode, separator, negative electrode, and separator, and then assembling them into cylindrical or square cells through a winding process. When winding the cells, the cathode and anode electrode sheets need to be correctly placed at both ends of the cell. The correct placement of the cell's tail is crucial to both performance and safety. Therefore, ensuring the correct tail placement of the cathode and anode electrode sheets is a very important process step in the cell manufacturing process.

[0092] In addition, for the ending of the anode and cathode pole pieces of the wound battery cell, the ending should be at the arc rather than on the large surface, because the corners of the arc do not participate in the lithium insertion and extraction reaction by default, that is, this part does not provide capacity. If the anode and cathode pole pieces end on the large surface, it is equivalent to losing part of the lithium ion capacity, but if they end on the arc, there will be no problem of low utilization.

[0093] However, the related art lacks a method for detecting the end position of the battery cell.

[0094] Based on this, an embodiment of the present application provides a method for detecting a tail position. After responding to a tail position detection instruction for a target battery cell, the method obtains internal structural images of the target battery cell from multiple different viewing angles, and then analyzes the internal structural images of the target battery cell from different viewing angles to determine the tail position detection result of the target battery cell. Since the internal structural image can reflect the density distribution and composition inside the target battery cell, the internal structural image of the target battery cell can be used to detect whether the tail position of the electrode inside the target battery cell is correct. In addition, by determining the tail position detection result using the internal structural images of the target battery cell from multiple different viewing angles, it is possible to reduce errors that may exist in a single view and reduce the possibility of misjudgment, thereby improving the accuracy of the tail position detection result.

[0095] Of course, it should be understood that the technical effects achievable by the tail position detection method provided in the embodiments of the present application are not limited to these, and other technical effects can also be achieved. For example, by detecting the tail position of the target battery cell's internal structure image, non-destructive testing of the target battery cell is achieved, and the efficiency of battery cell testing is improved. The specific technical effects achievable in the embodiments of the present application can be found in the following embodiments.

[0096] For ease of explanation, the following embodiments are described using a computer device according to one embodiment of the present application as the execution subject. The computer device is used to perform functional configuration on a target device. The computer device may be a server, and its internal structure diagram may be shown in FIG1 . The computer device includes a processor, a memory, an input / output interface (I / O), and a communication interface. The processor, the memory, and the I / O interface are connected via a system bus, and the communication interface is connected to the system bus via the I / O interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store the end position detection data of the device. The I / O interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, the steps of the end position detection method provided in any of the following embodiments of the present application are implemented.

[0097] Those skilled in the art will understand that the structure shown in FIG1 is merely a block diagram of a portion of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.

[0098] In an exemplary embodiment, as shown in FIG2 , a method for detecting a tail position is provided, the method comprising the following steps:

[0099] S201 , in response to a target battery cell end position detection instruction, acquiring internal structure images of the target battery cell at multiple different viewing angles.

[0100] The target battery cell may be any battery cell that needs to undergo final position detection after the winding process; the internal structure image of the target battery cell may reflect density distribution information, structural information, thickness information, and the like inside the target battery cell.

[0101] The internal structure image can be obtained after capturing the target battery cell by an image acquisition device; wherein the image acquisition device may include an X-ray (X-Ray) device, a magnetic resonance imaging device, an ultrasonic imaging device, an optical coherence tomography device, and the like; optionally, the internal structure image captured by the X-Ray device may be an X-ray image, the internal structure image captured by the magnetic resonance imaging device may be a magnetic resonance image, or a two-dimensional or three-dimensional grayscale image; the image captured by the ultrasonic imaging device may be an ultrasonic image, or a two-dimensional or three-dimensional grayscale image; the image captured by the optical coherence tomography device may be an optical coherence tomography image, or a high-resolution two-dimensional or three-dimensional image.

[0102] After the target battery cell undergoes the winding process on the battery production line, it reaches the finishing position detection process. The finishing position detection process may include a position sensor. After the position sensor senses that the target battery cell has reached the finishing position detection process, the finishing position detection instruction of the target battery cell may be sent to the computer device; or the target battery cell may be directly placed at a preset finishing position detection position, and the user triggers the finishing position detection instruction of the target battery cell on the computer device, and the computer device responds to the finishing position detection instruction triggered by the user.

[0103] The end position detection instruction may carry identification information of the target battery cell; the identification information may uniquely represent the target battery cell.

[0104] After responding to the tail position detection instruction, the computer device can obtain internal structure images of the target battery cell from multiple different viewing angles. The internal structure images of the target battery cell from multiple different viewing angles can be obtained by using an image acquisition device to obtain the internal structure images of the target battery cell from different angles. For example, by adjusting the acquisition angle of the image acquisition device, the internal structure images of the target battery cell from multiple different viewing angles can be obtained.

[0105] Optionally, the tail position detection instruction may carry internal structure images of the target battery cell at multiple different viewing angles, and the internal structure images of the target battery cell at multiple different viewing angles may be directly acquired from the tail position detection result.

[0106] It should be noted that the internal structure image can be a 16-bit, 24-bit or 32-bit image with depth information; the number of bits of the internal structure image can be determined according to actual needs, and the embodiments of the present application do not limit this.

[0107] S202: Analyze each internal structure image to determine a detection result of the end position of the target battery cell.

[0108] Among them, the end position detection results of the target battery cell include normal and abnormal.

[0109] The tail position detection result of the target battery cell can be determined by a preset detection model; specifically, each internal structure image is input into the detection model, and each internal structure image is analyzed by the detection model to determine the tail position detection result of the target battery cell.

[0110] Optionally, each internal structure image can also be compared with the corresponding standard image. If there is an internal structure image that is consistent with the standard image, the end position detection result of the target battery cell is determined to be normal; if each internal structure image is inconsistent with the corresponding standard image, the end position detection result of the target battery cell is determined to be abnormal.

[0111] Each view corresponds to a standard image, and the standard image may be an image showing a normal cell end position under the corresponding view.

[0112] It should be noted that if the detection result of the tail position of the target battery cell is normal, it means that the tail position of the electrode of the target battery cell is at the corner of the target battery cell. If the detection result of the tail position of the target battery cell is abnormal, it means that the tail position of the electrode of the target battery cell is on the large surface of the target battery cell. The large surface of the target battery cell refers to the surface with the largest area of ​​the battery cell, and the large surface can be the front or back of the battery cell.

[0113] Since the target battery cell includes a cathode electrode and an anode electrode, the end position detection of the target battery cell may include cathode end position detection and / or anode end position detection of the target battery cell, and the end position detection result of the target battery cell includes cathode end position detection result and / or anode end position detection result.

[0114] The end position detection result of the target battery cell may also include the end position detection result of the target battery cell as a whole, that is, when the end position detection result of the target battery cell is abnormal, there is an abnormality in the end position of the cathode and / or anode of the target battery cell; when the end position detection structure of the target battery cell is normal, the end positions of the cathode and anode of the target battery cell are both normal.

[0115] In the tail position detection method provided in an embodiment of the present application, in response to a tail position detection instruction of a target battery cell, internal structure images of the target battery cell at multiple different viewing angles are obtained, and then each internal structure image is analyzed to determine the tail position detection result of the target battery cell. In this method, the internal structure image can reflect the density distribution and composition inside the target battery cell. Therefore, the internal structure image of the target battery cell can be used to detect whether the tail position of the electrode inside the target battery cell is correct; and the tail position inside the target battery cell is detected by the internal structure image without destroying the integrity of the target battery cell, thereby achieving non-destructive detection of the tail position of the target battery cell; in addition, by determining the tail position detection result through the internal structure images of the target battery cell at multiple different viewing angles, it is possible to reduce the error that may exist in a single view, reduce the possibility of misjudgment, and improve the accuracy of the tail position detection result.

[0116] The multiple viewing angles of the target battery cell may include the overall viewing angle seen when observing the battery cell perpendicular to the large surface of the target battery cell. Therefore, the internal structure image includes the overall internal structure image of the target battery cell. In an exemplary embodiment, the internal structure images of the target battery cell at multiple different viewing angles are obtained, including: obtaining the overall internal structure image acquired by a first image acquisition device, and the acquisition viewing angle of the first image acquisition device covers the entire surface of the target battery cell.

[0117] The entire surface of the target battery cell can represent the large surface of the target battery cell. The acquisition perspective of the acquisition source of the first image acquisition device covers the large surface of the target battery cell. The first image acquisition device can directly capture the large surface of the target battery cell through the acquisition source, and determine the image of the target battery cell captured by the first image acquisition device as the overall internal structure image.

[0118] Optionally, the computer device may send an acquisition instruction to the first image acquisition device. After receiving the acquisition instruction, the first image acquisition device acquires the overall internal structure image of the target battery cell and sends the acquired overall internal structure image to the computer device.

[0119] The first image acquisition device may be an X-ray device, a magnetic resonance imaging device, an ultrasonic imaging device, an optical coherence tomography device, and the like.

[0120] Taking the first image acquisition device as an X-Ray device as an example, the overall internal structure image acquired by the X-Ray device can be an X-ray image; as shown in Figure 3, Figure 3 is a structural schematic diagram of the X-Ray device acquiring the overall internal structure image, wherein 301 represents the target battery cell, 302 represents the large surface of the target battery cell, 303 represents the first X-Ray device, 304 represents the acquisition source of the first X-Ray device, and 305 represents the flat-panel detector of the first X-Ray device.

[0121] The X-ray device sends X-rays through the collection source. The X-rays penetrate the target cell from the large surface and reach the flat-panel detector in the X-ray device. The flat-panel detector displays the image based on the amount of radiation received, and obtains the overall internal structure image. Among them, the X-rays emitted by the X-ray device have strong penetrating properties. Due to the difference in absorption degree caused by factors such as the density and thickness of the object, the amount of radiation reaching the flat-panel detector in the X-ray device is different and displayed as an image; its attenuation formula is formula (1). I = I0e -μd (1)

[0122] Among them, is the intensity of the ray after penetrating the object, is the intensity of the incident ray, is the ray attenuation coefficient, and is the thickness of the object.

[0123] In the end position detection method provided in the embodiments of the present application, a first image acquisition device captures an overall internal structure image, and the first image acquisition device's capture angle of view covers the entire surface of the target battery cell. In this method, the first image acquisition device's capture angle of view covers the entire surface of the target battery cell, so that the captured overall internal structure image can represent all density information within the target battery cell, facilitating a comprehensive analysis of the target battery cell's internal structure, thereby improving the accuracy of the end position detection results.

[0124] Since the end position of the electrode of the target cell should be at the corner of the target cell, the internal structure image of the corner of the target cell can be obtained, and the end position of the electrode can be detected through the internal structure image of the corner.

[0125] Therefore, the internal structure image may include an internal structure image of a corner of the target battery cell. In an exemplary embodiment, as shown in FIG4 , obtaining the internal structure images of the target battery cell at multiple different viewing angles includes the following steps:

[0126] S401, obtaining a first internal structure image captured by a second image acquisition device and a second internal structure image captured by a third image acquisition device; the second image acquisition device has a capture angle covering an area where the first corner and the second corner of the target battery cell are located; the third image acquisition device has a capture angle covering an area where the third corner and the fourth corner of the target battery cell are located.

[0127] The second image acquisition device and the third image acquisition device may be the same as the first image acquisition device.

[0128] Take the second image acquisition device and the third image acquisition device as X-Ray devices as an example; the large surface direction of the target battery cell includes 4 corners, among which two corners can be used as the first corner and the second corner, and the other two corners can be used as the third corner and the fourth corner; as shown in Figure 5, Figure 5 is a structural schematic diagram of the X-Ray device acquiring the second internal structure image and the third internal structure image; 501 can be the area where the first corner is located, 502 can be the area where the second corner is located, 503 can be the area where the third corner is located, and 504 can be the area where the fourth corner is located; the second X-Ray device 505 can be fixedly set vertically above the area 501 where the first corner is located and the area 502 where the second corner is located, and the third X-Ray device 506 can be fixedly set vertically above the area 503 where the third corner is located and the area 504 where the fourth corner is located.

[0129] The capture angle of the second image capture device can cover the area where the first corner and the second corner of the target battery cell are located; the capture angle of the third image capture device can cover the area where the third corner and the fourth corner of the target battery cell are located.

[0130] After the computer device responds to the target battery cell's end position detection instruction, it can send an acquisition instruction to the second image acquisition device and the third image acquisition device; after receiving the acquisition instruction, the second image acquisition device acquires the area where the first corner is located and the area where the second corner is located to obtain a first internal structure image, and sends the first internal structure image to the computer device; after receiving the acquisition instruction, the third image acquisition device acquires the area where the third corner is located and the area where the fourth corner is located to obtain a second internal structure image, and sends the second internal structure image to the computer device.

[0131] Among them, after the second image acquisition device and the third image acquisition device acquire the first internal structure image and the second internal structure image, they can process the first internal structure image and the second internal structure image, and send the processed first internal structure image and the second internal structure image to the computer device for use in detecting the end position of the target battery cell. Taking the first internal structure image as an example, the first internal structure image acquired by the first image acquisition device may also include other areas besides the first corner area and the second corner area. The other areas can be deleted, and only the area where the first corner and the area where the second corner are located are retained; the processing principle of the second internal structure image is the same as that of the first internal structure image, and the embodiments of this application will not be repeated here.

[0132] S402: Merge the first internal structure image and the second internal structure image to obtain a corner internal structure image.

[0133] After receiving the first internal structure image and the second internal structure image, the computer device merges the first internal structure image and the second internal structure image into one image to obtain a corner internal structure image.

[0134] Among them, the way to merge the first internal structure image and the second internal structure image can be to combine the first internal structure image and the second internal structure image into a corner internal structure image, and the corner internal structure image includes the first internal structure image and the second internal structure image; that is, the corner internal structure image includes the area where the first corner is located, the area where the second corner is located, the area where the third corner is located, and the area where the fourth corner is located.

[0135] In the end position detection method provided in the embodiment of the present application, a first internal structure image captured by a second image acquisition device and a second internal structure image captured by a third image acquisition device are obtained, and the first internal structure image and the second internal structure image are merged to obtain a corner internal structure image; wherein, the capture angle of the second image acquisition device covers the area where the first corner and the second corner are located on the target battery cell; and the capture angle of the third image acquisition device covers the area where the third corner and the fourth corner are located on the target battery cell. In this method, since the end position of the target battery cell is detected in order to detect whether the end position of the battery cell is at the corner of the target battery cell, the end position of the corner internal structure image is detected in a targeted manner by capturing an image of the local corner area of ​​the target battery cell, thereby improving the accuracy and efficiency of the end position detection.

[0136] The above embodiment describes how to acquire the overall internal structure image and corner internal structure image of the target battery cell. The following embodiment describes how to detect the battery cell end position using the overall internal structure image and corner internal structure image.

[0137] In an exemplary embodiment, each internal structure image includes an overall internal structure image and a corner internal structure image of the target battery cell. As shown in FIG6 , analyzing each internal structure image to determine the end position detection result of the target battery cell includes the following steps:

[0138] S601 , performing image enhancement processing on the overall internal structure image and the corner internal structure image respectively to obtain an enhanced overall internal structure image and an enhanced corner internal structure image.

[0139] Due to factors such as the image acquisition device itself and the environment, the captured overall internal structure image and corner internal structure image of the target battery cell may contain noise, errors, and other problems. Therefore, before performing tail position detection on the target battery cell using the overall internal structure image and corner internal structure image, the overall internal structure image and the corner internal structure image can be subjected to image enhancement processing. The overall internal structure image after image enhancement processing is determined as the enhanced overall internal structure image, and the corner internal structure image after image enhancement processing is determined as the enhanced corner internal structure image.

[0140] Among them, the image enhancement processing method for the overall internal structure image and the corner internal structure image can include at least one of image transformation and correction, histogram equalization, sharpening, noise removal, scale transformation, contrast enhancement, and region of interest extraction.

[0141] S602 : Determine a detection result of a tail position of a target battery cell according to the overall internal structure image, the enhanced overall internal structure image, and the enhanced corner internal structure image.

[0142] In one embodiment, the overall internal structure image, the enhanced overall internal structure image and the enhanced corner internal structure image can be input into a target detection model, and the overall internal structure image, the enhanced overall internal structure image and the enhanced corner internal structure image can be analyzed by the target detection model to obtain the position of the battery cell tail position output by the target detection model on the target battery cell.

[0143] If the cell end position is at the corner of the target cell, the target cell end position detection result is determined to be normal; if the cell end position is not at the corner of the target cell, the target cell end position detection result is determined to be abnormal.

[0144] In another embodiment, the overall internal structure image, the enhanced overall internal structure image and the enhanced corner internal structure image can also be input into the target detection model respectively to obtain the detection results of each internal structure image; the detection results include whether the end position of the battery cell is at the corner of the target battery cell.

[0145] If there are two or more detection results that both indicate that the cell end position is at the corner of the target cell, the target cell end position detection result is determined to be normal; otherwise, the target cell end position detection result is abnormal.

[0146] In the method for detecting the end position provided in an embodiment of the present application, image enhancement processing is performed on the overall internal structure image and the corner internal structure image, respectively, to obtain an enhanced overall internal structure image and an enhanced corner internal structure image. The end position detection result of the target battery cell is determined based on the overall internal structure image, the enhanced overall internal structure image, and the enhanced corner internal structure image. In this method, image enhancement processing is performed on the overall internal structure image and the corner internal structure image, which can highlight important features in the overall internal structure image and the corner internal structure image. Therefore, by detecting the end position using the overall internal structure image and the enhanced overall internal structure image and the enhanced corner internal structure image after enhancement processing, the accuracy of the end position detection result can be improved. Furthermore, by performing comprehensive detection of the end position using internal structure images of multiple dimensions, the accuracy of end position detection is further improved.

[0147] Since non-battery cell areas may be captured when capturing internal structure images, in order to improve detection efficiency and accuracy, the area image required for detecting the end position can be first extracted from the overall internal structure image and the corner internal structure image, and then the extracted area image can be image processed. The following is a detailed description of this through an embodiment. In one embodiment, as shown in Figure 7, image enhancement processing is performed on the overall internal structure image and the corner internal structure image respectively to obtain an enhanced overall internal structure image and an enhanced corner internal structure image, including the following steps:

[0148] S701, determining edge pixels of a target image according to the grayscale value of each pixel in the target image; the target image is an overall internal structure image or a corner internal structure image.

[0149] Since the grayscale values ​​between the battery cell area and the non-battery cell area are greatly different, the edge pixels of the target image can be determined according to the grayscale value of each pixel in the target image.

[0150] For example, pixels whose grayscale value differences between adjacent pixels in the target image are greater than a preset difference threshold are determined as edge pixels, thereby determining all edge pixels of the target image.

[0151] S702 : Acquire a region of interest image from the target image according to edge pixels of the target image and a preset outward expansion offset.

[0152] The area enclosed by the edge pixel points of the target image can be determined as the area where the battery cells are located in the target image; an area image enclosed by expanding outward by an offset with the edge pixel point as the center is intercepted from the target image, and this area image is determined as the area of ​​interest image.

[0153] The target image is taken as the overall internal structure image for explanation, as shown in Figure 8, which is a structural diagram of extracting the region of interest image, 801 is the overall internal structure image, 802 is the edge pixel point of the overall internal structure image 801, and then the edge pixel point is expanded outward by the outward expansion offset to obtain the region of interest image 803; it should be noted that the overall internal structure image in Figure 8 only illustrates the outline, and does not illustrate the internal structure distribution composed of different gray levels.

[0154] S703: Perform image processing on the region of interest image to obtain an enhanced internal structure image of the target image.

[0155] The region of interest image after image processing is determined as the enhanced internal structure image of the target image.

[0156] In one embodiment, the image of the region of interest can be processed according to a preset image processing model; specifically, the image of the region of interest is input into the image processing model, and the image processing model is used to process the region of interest to obtain an enhanced internal structure image of the target image.

[0157] In another embodiment, as shown in FIG9 , performing image processing on the region of interest image to obtain an enhanced internal structure image of the target image includes the following steps:

[0158] S901 , performing logarithmic domain conversion on the region of interest image to obtain a logarithmic domain internal structure image.

[0159] Since the logarithmic transformation can map smaller grayscale values ​​in the image to larger grayscale values, it can expand the low grayscale value part of the image, improve the contrast of the darker area, and make the details in the image more prominent.

[0160] Therefore, the image of the region of interest can be converted into the logarithmic domain, and the image of the region of interest after the logarithmic domain conversion can be determined as the logarithmic domain internal structure image; specifically, the logarithm of the grayscale value corresponding to each pixel point in the image of the region of interest can be directly calculated, and then the grayscale value after the logarithm calculation can be determined as the grayscale value of the logarithmic domain internal structure image, thereby obtaining the logarithmic domain internal structure image.

[0161] S902 , performing bilateral filtering on the logarithmic domain internal structure image to obtain a filtered internal structure image.

[0162] Bilateral filtering can combine the spatial proximity and pixel value similarity of an image, while taking into account the spatial information and grayscale similarity of the image, and can perform filtering while maintaining image details.

[0163] Therefore, the logarithmic domain internal structure image may be subjected to bilateral filtering processing, and the logarithmic domain internal structure image after the bilateral filtering processing may be determined as the filtered internal structure image.

[0164] The method of performing bilateral filtering on the logarithmic domain internal structure image can be to take any pixel point in the logarithmic domain internal structure image as the target pixel point, and obtain the template area of ​​the target pixel point with the target pixel point as the center; and filter the target pixel point through the template area to obtain the pixel information of the target pixel point after filtering; and determine the filtered internal structure image based on the pixel information of each target pixel point after filtering.

[0165] The template area image of the target pixel point can be obtained by taking the target pixel point as the center and determining the area within the vicinity of the target pixel point as the template area of ​​the target pixel point; for example, the length and width of the template can be determined according to the size of the internal structure image of the logarithmic domain.

[0166] For example, if the template area is a 7*7 pixel template, 7*7 neighboring pixels adjacent to the target pixel are determined as pixels within the template area.

[0167] Then, the target pixel is filtered based on the pixel points within the template area to obtain the pixel grayscale value of the target pixel after filtering. The filtered internal structure image is determined based on the pixel grayscale value of each target pixel. The method for filtering the target pixel can be shown in formula (2) and formula (3). I = I0e -μd (2) I = I0e -μd (3)

[0168] Among them, represents the spatial position of the target pixel point, represents the corresponding template area, is the spatial position of any pixel point in the template area, represents the grayscale value of the target pixel point, represents the grayscale value of any pixel point in the template area, represents the absolute value of the grayscale difference between the target pixel point and a pixel point in the template area, represents the Gaussian kernel function of the geometric proximity relationship, represents the Gaussian kernel function of the grayscale similarity relationship, represents the spatial Euclidean distance between the target pixel point and a pixel point in the template area, represents the grayscale value of the target pixel point after filtering, and represents the normalization coefficient.

[0169] This embodiment uses bilateral filtering to filter the image, which can achieve the effect of maintaining edges, reducing noise and smoothing. As shown in formula (2), the intensity of a pixel is represented by a weighted average method based on Gaussian distribution, while considering the Euclidean distance of the pixel and the radiation difference in the pixel range domain.

[0170] It should be noted that since the Gaussian function is a form of statistical function, its function shape is a normal distribution centered on the expected value and with the standard deviation as the confidence interval. The size of the standard deviation can determine the validity of the function range. Therefore, when calculating the Gaussian kernel function of the geometric proximity relationship and the Gaussian kernel function of the grayscale similarity relationship, the spatial standard deviation parameters of the geometric proximity relationship and the intensity standard deviation parameters of the grayscale similarity relationship are first determined based on each pixel point in the template area. Then, the Gaussian kernel function of the geometric proximity relationship is constructed based on the spatial standard deviation parameters, and the Gaussian kernel function of the grayscale similarity relationship is constructed based on the intensity standard deviation parameters.

[0171] S903: Perform equalization processing on the filtered internal structure image to obtain an enhanced internal structure image of the target image.

[0172] Performing histogram equalization on an image can enhance the global contrast of the image. Therefore, the filtered internal structure image can be subjected to histogram equalization, and the filtered internal structure image after the histogram equalization process is determined as the enhanced internal structure image of the target image.

[0173] The method for equalizing the filtered internal structure image can be to determine the number of pixels at each grayscale level in the filtered image based on the grayscale histogram of the filtered image; map each grayscale value in the filtered image based on the number of pixels to obtain a new grayscale value after equalization; and construct an enhanced internal structure image based on each new grayscale value. This is shown in formula (4). I = I0e -μd (4)

[0174] Among them, represents the grayscale value of the pixel point with grayscale value in the filtered internal structure image after equalization, represents the grayscale transformation function, represents the grayscale value of level, represents the number of pixels with grayscale of in the filtered internal structure image, represents the grayscale level range of the filtered internal structure image, and represents the number of rows and columns of pixels in the filtered internal structure image respectively.

[0175] In this embodiment, the region of interest image is transformed into a logarithmic domain to obtain a logarithmic domain internal structure image, and the logarithmic domain internal structure image is subjected to bilateral filtering to obtain a filtered internal structure image. The filtered internal structure image is then equalized to obtain an enhanced internal structure image of the target image. In this method, the logarithmic domain transformation can improve the image's detail sensitivity, the bilateral filtering can effectively reduce the image's noise and improve the image quality, and the equalization process can enhance the image's contrast, making the grayscale changes in the image more obvious. Therefore, by performing this series of processing on the region of interest image, a clearer and more accurate internal structure image can be obtained, providing more accurate data support for subsequent detection of the end position.

[0176] In the end position detection method provided in an embodiment of the present application, the edge pixels of the target image are determined based on the grayscale value of each pixel in the target image; the target image is an overall internal structure image or a corner internal structure image; and based on the edge pixels of the target image and a preset outward expansion offset, a region of interest image is obtained from the target image, and then image processing is performed on the region of interest image to obtain an enhanced internal structure image of the target image. In this method, the region of interest images in the overall internal structure image and the corner internal structure image are first extracted, and then image processing is performed on the region of interest images. This reduces the computational complexity of enhancing the overall internal structure image or the corner internal structure image and improves image processing efficiency. Furthermore, when extracting the region of interest image, the region of interest is determined based on the edge pixels and the outward expansion offset, making the extracted region of interest more accurate and effective.

[0177] The above embodiment describes how to obtain an enhanced internal structure image. The following describes how to obtain the tail position detection result of the target battery cell through an embodiment. In an exemplary embodiment, the tail position detection result of the target battery cell is determined based on the overall internal structure image, the enhanced overall internal structure image and the enhanced corner internal structure image, including: inputting the overall internal structure image, the enhanced overall internal structure image and the enhanced corner internal structure image into the position detection model to obtain the tail position detection result of the target battery cell.

[0178] Among them, the position detection model can be a neural network model that is pre-trained based on a large number of historical overall internal structure images, historical enhanced overall internal structure images and historical enhanced corner internal structure images, and is specifically used to perform tail position detection on the target battery cell.

[0179] After obtaining the overall internal structure image, enhanced overall internal structure image and enhanced corner internal structure image of the target battery cell, the computer device can input the overall internal structure image, enhanced overall internal structure image and enhanced corner internal structure image of the target battery cell into a pre-trained position detection model. After analyzing the overall internal structure image, enhanced overall internal structure image and enhanced corner internal structure image through the position detection model, the position detection model can directly output the finishing position detection result of the target battery cell.

[0180] Optionally, when the overall internal structure image, enhanced overall internal structure image and enhanced corner internal structure image of the target battery cell are input into a pre-trained position detection model, the position detection model can output a probability value that the corner position of the target battery cell is normal; when the probability value is greater than a preset probability threshold, the tail position detection result of the target battery cell is determined to be a normal tail position; when the probability value is less than or equal to the preset probability threshold, the tail position detection result of the target battery is determined to be an abnormal tail position.

[0181] Optionally, some basic neural network models may be trained to obtain a position detection model. For example, the basic neural network models include but are not limited to deep learning network models, deep convolutional neural network models, residual neural network (ResNet) models, etc.

[0182] In the tail position detection method provided in the embodiments of the present application, the overall internal structure image, the enhanced overall internal structure image, and the enhanced corner internal structure image are all input into a position detection model to obtain the tail position detection result of the target battery cell. In this method, the tail position detection result of the target battery cell is directly determined using a pre-trained position detection model, improving the efficiency and accuracy of tail position detection for the target battery cell.

[0183] In the above embodiment, the position detection model may include a depthwise separable convolutional network and an output network. In an exemplary embodiment, as shown in FIG10 , the overall internal structure image, the enhanced overall internal structure image, and the enhanced corner internal structure image are all input into the position detection model to obtain the tail position detection result of the target battery cell, including the following steps:

[0184] S1001, multi-scale feature extraction is performed on the overall internal structure image, enhanced overall internal structure image and enhanced corner internal structure image through a deep separable convolutional network, and overall feature images, enhanced overall feature images and enhanced corner feature images at multiple different scales are obtained respectively.

[0185] A depthwise separable convolutional network may include multiple depthwise separable convolutional layers, each of which may include channel-wise convolution and point-wise convolution.

[0186] The overall internal structure image, the enhanced overall internal structure image and the enhanced corner internal structure image can be input into the depthwise separable convolutional network respectively to obtain overall feature images at multiple different scales, enhanced overall feature images at multiple different scales and enhanced corner feature images at multiple different scales respectively.

[0187] For any of the internal structure images, namely, the overall internal structure image, the enhanced overall internal structure image, and the enhanced corner internal structure image, multi-scale feature extraction can be performed on the internal structure image through multiple depth-wise separable convolutional layers to obtain feature images at multiple different scales.

[0188] As shown in Figure 11, Figure 11 is a structural diagram of a depth-wise separable convolutional network, which is illustrated by taking the input of the overall internal structure image and the depth-wise separable convolutional network including three depth-wise separable convolutional layers as an example. Figure 11 illustrates the overall feature images obtained at three different scales: the first overall feature image, the second overall feature image, and the third overall feature image.

[0189] It should be noted that there is no limit on the number of depthwise separable convolutional layers and the number of feature extraction scales in the depthwise separable convolutional network, and they can be set according to actual conditions.

[0190] S1002: Input each overall feature image, each enhanced overall feature image, and each enhanced corner feature image into the output network to obtain a tail position detection result.

[0191] The overall feature images at each scale, the enhanced overall feature images at each scale, and the enhanced corner feature images at each scale are input into the output network. The output network analyzes each overall feature image, each enhanced overall feature image, and each enhanced corner feature image to obtain the tail position detection result output by the output network.

[0192] In the end position detection method provided in the embodiments of the present application, a depthwise separable convolutional network is used to perform multi-scale feature extraction on the overall internal structure image, the enhanced overall internal structure image, and the enhanced corner internal structure image, respectively. This results in overall feature images, enhanced overall feature images, and enhanced corner feature images at multiple scales. Each of these overall feature images, enhanced overall feature images, and enhanced corner feature images is then input into an output network to obtain the end position detection result. This method determines the end position detection result of the target battery cell using feature images at different angles and multiple scales, thereby improving the accuracy of the end position detection result.

[0193] The above is an explanation of the application process of the position detection model. The following is an explanation of the construction process of the position detection model through an embodiment. In an exemplary embodiment, the embodiment includes: obtaining an initial position detection model and a training sample set, iteratively training the initial position detection model through the training sample set until the initial position detection model converges to obtain a position detection model.

[0194] Among them, considering that industrial target detection has high requirements for algorithm efficiency, YOLOv5 can be selected as the benchmark model, which has the advantages of fast detection speed and high accuracy, as shown in Figure 12. Figure 12 is a structural diagram of the YOLOv5 benchmark model; among them, the YOLOv5 benchmark model consists of a backbone network, a neck network and a head network; the backbone network generates three feature images of different scales based on the input image, and the neck network performs upward fusion and downward fusion of features on the three feature images of different scales output by the backbone network, thereby combining shallow graphic features with deep semantic features to obtain three more complete feature maps; the head network performs 1*1 convolution on the three feature maps output by the neck network to obtain the final three feature maps of the input image; the YOLOv5 model can subsequently be detected through the three feature maps generated by the head network.

[0195] Specifically, the attention mechanism can extract regional images or features of the input image; the convolution block can extract features and form a feature image. The convolution block can be composed of a convolution layer, a batch normalization layer and an activation function; feature extraction block 1, feature extraction block 2 and feature extraction block 3 are feature extraction modules of different scales, and spatial pyramid pooling can convert feature maps of non-fixed scales into a unified scale.

[0196] However, considering the limited memory of industrial equipment, the YOLOv5 benchmark model can be lightweighted to reduce the model size and the memory usage time at runtime. Therefore, the backbone model of YOLOv5 can be replaced with the MobileNetV3 lightweight neural network, which greatly reduces the model size and computational complexity while ensuring high accuracy. The MobileNetV3 model consists of a first convolutional block, a second convolutional block, an average pooling layer, a third convolutional block and a fully connected layer. The first convolutional block includes a convolutional layer, a batch normalization layer and an activation function. The first convolutional block can extract the features of the input image. The second convolutional block includes multiple convolutional layers, a lightweight attention model, a nonlinear activation function, etc. The third convolutional block includes a fully connected layer, a batch normalization layer and an activation function. The convolutional layer in the MobileNetV3 model can be a depth-wise separable convolutional layer. Among them, the depth-wise separable convolutional layer can decompose the standard convolutional layer into channel-by-channel convolution and point-by-point convolution, reducing the computational complexity and parameters of the convolution.

[0197] Based on this, according to the idea of ​​lightweight MobileNetV3 model, the backbone network of YOLOv5 benchmark model can be modified, and the idea of ​​depthwise separable convolutional layer can be applied to the YOLOv5 benchmark model to determine the initial position detection model, as shown in Figure 13, which is a structural diagram of the initial position detection model.

[0198] It should be noted that the backbone network in Figure 13 includes not only channel-by-channel convolution and point-by-point convolution, but also other network layers. The structure of the initial position detection model in the embodiment of the present application is only an example; and the network structure of the YOLOv5 benchmark model can also be modified by other lightweight ideas of the MobileNetV3 model, which will not be repeated here in the embodiment of the present application.

[0199] The training sample set can be obtained by first obtaining the historical overall internal structure image and the historical corner internal structure image of the historical battery cell; then performing image enhancement processing on the historical overall internal structure image and the historical corner internal structure image to obtain the historical enhanced overall internal structure image and the historical enhanced corner internal structure image; then annotating the historical enhanced overall internal structure image, the historical enhanced corner internal structure image and the historical overall internal structure image; and determining the annotated historical enhanced overall internal structure image, the historical enhanced corner internal structure image and the historical overall internal structure image as the training sample set.

[0200] It should be noted that the method of performing image enhancement processing on the historical overall internal structure image and the historical corner internal structure image is the same as the method of performing enhancement processing on the overall internal structure image and the corner internal structure image in the above embodiment, and this embodiment will not be repeated here.

[0201] [Corrected on 11.09.2024 according to Rule 26] The historical enhanced overall internal structure image, the historical enhanced corner internal structure image and the historical overall internal structure image can be annotated using the LabelImg annotation software. The annotation results can include whether there is a tail position in the historical enhanced overall internal structure image, the historical enhanced corner internal structure image and the historical overall internal structure image, and whether the tail position is normal. After the historical enhanced overall internal structure image, the historical enhanced corner internal structure image and the historical overall internal structure image are annotated using the LabelImg annotation software, the annotation results can also be modified by review.

[0202] Based on the initial position detection model and training sample set obtained above, the training sample set can be input into the initial position detection model, and the initial position detection model can be trained by the training sample set until the initial position detection model reaches the preset convergence condition, and the initial position detection model that reaches the convergence condition is determined as the position detection model.

[0203] Among them, the initial weights in the initial position detection model can be public model pre-training weights, which are features that have been learned on large-scale data sets. When the initial position detection model is trained, some weights are frozen and the remaining weights are fine-tuned or modified. In this way, freezing some weights can retain the features learned in the pre-training stage and speed up model training; fine-tuning or modifying some weights can make the model better adapt to specific tasks or data sets.

[0204] The convergence condition of the initial position detection model may be that both the average accuracy mean and the F1 score of the initial position detection model reach a preset threshold. For example, during the iterative training of the initial position detection model, the initial position detection model after each iterative training is tested based on a test data set to determine the average accuracy mean and the F1 score of the initial position detection model. If the average accuracy mean reaches an accuracy threshold and the F1 score reaches a preset score threshold, the initial position detection model is determined as the position detection model.

[0205] The above embodiment describes how the target cell's tail position detection result is determined using a position detection model. The following describes another method for determining the tail position detection result using an embodiment. In one exemplary embodiment, as shown in FIG14 , the tail position detection result of the target cell is determined based on the overall internal structure image, the enhanced overall internal structure image, and the enhanced corner internal structure image, including the following steps:

[0206] S1401 , determining a first probability value that a tail position of a target battery cell is normal based on the overall internal structure image and a preset standard overall internal structure image.

[0207] The standard overall internal structure image may include an overall internal structure image with a normal tail position.

[0208] The overall internal structure image is compared with a standard overall internal structure image to obtain a similarity between the overall internal structure image and the standard overall internal structure image, and the similarity is determined as a first probability value that the target battery cell tail position is normal.

[0209] Optionally, the situation where the tail position of the target battery cell is normal may include multiple situations, therefore, the standard overall internal structure image may be multiple standard overall internal structure images; the similarity between the overall internal structure image and each standard overall internal structure image may be obtained; the maximum similarity may be determined as the first probability value that the tail position of the target battery cell is normal.

[0210] Among them, the calculation method of the similarity between the overall internal structure image and the standard overall internal structure image may include: calculating the cosine similarity between the overall internal structure image and the standard overall internal structure image based on the pixel grayscale value of the overall internal structure image and the pixel grayscale value of the standard overall internal structure image; determining the first probability value based on the cosine similarity; and determining the first probability value based on the principle that the greater the cosine similarity, the higher the probability.

[0211] For example, the computer device has multiple correspondences between cosine similarity ranges and probability values, and determines the probability value corresponding to the cosine similarity between the overall internal structure image and the standard overall internal structure image in the correspondence as the first probability value.

[0212] The similarity between the overall internal structure image and the standard overall internal structure image may also be calculated using methods such as Euclidean distance, mean square error, and correlation coefficient, which is not limited in this embodiment of the present application.

[0213] S1402 : Determine a second probability value that the tail position of the target battery cell is normal based on the enhanced overall internal structure image and a preset standard enhanced overall internal structure image.

[0214] S1403 : Determine a third probability value that the tail position of the target battery cell is normal based on the enhanced corner internal structure image and a preset standard enhanced corner internal structure image.

[0215] It should be noted that, in this embodiment, the second probability value of the normal tail position of the target battery cell is determined based on the enhanced overall internal structure image and the preset standard enhanced overall internal structure image, and the third probability value of the normal tail position of the target battery cell is determined based on the enhanced corner internal structure image and the preset standard enhanced corner internal structure image. The method is the same as the method of determining the first probability value of the normal tail position of the target battery cell based on the overall internal structure image and the preset standard overall internal structure image, and the embodiments of the present application will not be repeated here.

[0216] S1404: Determine a detection result of the end position of the target battery cell according to the first probability value, the second probability value, and the third probability value.

[0217] The detection results of the tail position of the target battery cell include normal and abnormal. The detection results of multiple views can be fused to obtain the detection results of the tail position of the target battery cell.

[0218] In one embodiment, the maximum probability value among the first probability value, the second probability value and the third probability value can be determined as the fourth probability value that the tail position of the target battery cell is normal, and then the fourth probability value is compared with a preset normal probability threshold. When the fourth probability value is greater than the normal probability threshold, the tail position detection result of the target battery cell is determined to be normal; when the fourth probability value is less than or equal to the normal probability threshold, the tail position detection result of the target battery cell is determined to be abnormal.

[0219] In another embodiment, as shown in FIG15 , determining the end position detection result of the target battery cell according to the first probability value, the second probability value, and the third probability value includes the following steps:

[0220] S1501: Determine a comprehensive probability value based on the first probability value, the second probability value, and the third probability value.

[0221] The comprehensive probability value can represent the probability value that the tail position of the target battery cell is normal.

[0222] The first probability value, the second probability value and the third probability value can be weightedly calculated to obtain a comprehensive probability value; specifically, the first weight of the overall internal structure image, the second weight of the enhanced overall internal structure image and the third weight of the enhanced corner internal structure image can be first obtained, and the corresponding first probability value, the second probability value and the third probability value can be weightedly calculated through the first weight, the second weight and the third weight, and the result obtained by the weighted calculation is determined as the comprehensive probability value.

[0223] Among them, the first weight, the second weight and the third weight can be determined based on historical experience; for example, the third weight of enhancing the corner internal structure image is greater than the second weight of enhancing the overall internal structure image, and the second weight of enhancing the overall internal structure image is greater than the first weight of the overall internal structure image; the sum of the first weight, the second weight and the third weight can be set equal to 1.

[0224] Optionally, the sum of the first probability value, the second probability value and the third probability value may be determined as the comprehensive probability value; or the maximum probability value among the first probability value, the second probability value and the third probability value may be determined as the comprehensive probability value.

[0225] S1502: When the comprehensive probability value is greater than the preset probability threshold, the tail position detection result is determined to be normal; when the comprehensive probability value is less than or equal to the probability threshold, the tail position detection result is determined to be abnormal.

[0226] The probability threshold may be determined based on historical experience, for example, the probability threshold may be 80%.

[0227] If the comprehensive probability value is greater than a preset probability threshold, it indicates that the probability that the tail position of the target battery cell is normal is relatively high, and the tail position detection result is determined to be normal.

[0228] If the comprehensive probability value is less than or equal to the probability threshold, it means that the probability that the tail position of the target battery cell is normal is small, and the tail position detection result is determined to be abnormal.

[0229] In this embodiment, a comprehensive probability value is determined based on the first, second, and third probability values. If the comprehensive probability value is greater than a preset probability threshold, the tail position detection result is determined to be normal; if the comprehensive probability value is less than or equal to the probability threshold, the tail position detection result is determined to be abnormal. In this method, a decision is made based on multiple probabilities indicating normal tail position detection, and the tail position detection result of the target battery cell is determined based on the resulting comprehensive probability value and the preset probability threshold, thereby improving the accuracy of tail position detection.

[0230] In the tail position detection method provided in the embodiment of the present application, a first probability value that the tail position of the target battery cell is normal is determined based on the overall internal structure image and a preset standard overall internal structure image, and a second probability value that the tail position of the target battery cell is normal is determined based on the enhanced overall internal structure image and the preset standard enhanced overall internal structure image. Then, a third probability value that the tail position of the target battery cell is normal is determined based on the enhanced corner internal structure image and the preset standard enhanced corner internal structure image. Finally, based on the first probability value, the second probability value, and the third probability value, the tail position detection result of the target battery cell is determined. In this method, the overall internal structure image, the enhanced overall internal structure image, and the enhanced corner internal structure image are respectively compared with the corresponding standard internal structure image to determine multiple probability values ​​that the tail position of the target battery cell is normal. Then, the multiple probability values ​​are fused for decision making to determine the tail position detection result of the target battery cell, thereby improving the comprehensiveness of the tail position detection and thus improving the effectiveness and accuracy of the tail position detection result.

[0231] In the above embodiment, taking the example that the tail position detection result of the target battery cell includes cathode tail position detection and anode tail position detection, then, based on the overall internal structure image and the preset standard overall internal structure image, a first probability value that the cathode and anode tail positions of the target battery cell are normal is determined, and based on the enhanced overall internal structure image and the preset standard enhanced overall internal structure image, a second probability value that the cathode and anode tail positions of the target battery cell are normal is determined, and then based on the enhanced corner internal structure image and the preset standard enhanced corner internal structure image, a third probability value that the cathode and anode tail positions of the target battery cell are normal is determined, and finally, based on the first probability value, second probability value and third probability value corresponding to the cathode and anode respectively, the cathode tail position detection result and the anode tail position detection result of the target battery cell are determined.

[0232] Taking the first probability value as an example, the standard overall internal structure image may include an overall internal structure image with a normal cathode end position and an overall internal structure image with a normal anode end position; then, based on the overall internal structure image and the overall internal structure image with a normal cathode end position, the first probability value of the cathode end position of the target battery cell being normal is determined, and based on the overall internal structure image and the overall internal structure image with a normal anode end position, the first probability value of the anode end position of the target battery cell being normal is determined.

[0233] During the winding process, the anode active material coating of the battery cell needs to be able to wrap around the cathode active material coating to prevent lithium deposition. Therefore, for wound batteries, the width of the anode should be wider than the width of the cathode, and the length of the anode should be longer than the length of the cathode. The OverHang detection method can be used to detect the end position of the battery cell electrode. That is, images of the four corners of the target battery cell are collected and the end position of the four corner images is detected.

[0234] The following is an explanation taking the cathode end position as an example, as shown in Figure 16, which is a schematic diagram of X-Ray imaging of the corner area of ​​a target battery cell. Figure (a) in Figure 16 represents a planar image of the corner area of ​​the target battery cell. Figure (a) in Figure 16 illustrates the situation where the cathode end position of the battery cell is at the corner. In this case, the winding of the anode and the cathode at the corner position is alternating in sequence, as shown in Figure (b) in Figure 16.

[0235] As shown in Figure 17, Figure 17 is a schematic diagram of X-Ray imaging of the corner area of ​​another target battery cell. Figure (a) in Figure 17 represents a planar image of the corner area of ​​the target battery cell. Figure (a) in Figure 17 shows the situation where the cathode end position of the battery cell is not at the corner. In this case, there will be two adjacent anode lines when the anode and cathode are wound at the corner position, as shown in Figure 17 (b).

[0236] As shown in FIG18 , FIG18 is a schematic diagram of the overall internal structure image of the battery cell, and the overall internal structure image shows that the cathode end position of the battery cell is normal.

[0237] It should be noted that Figures 8, 16-18 in the embodiments of the present application are all examples made with black and white lines. In actual applications, the overall internal structure image of the battery cell and the planar image of the corner area of ​​the battery cell can be composed of different grayscale levels; for example, in Figures 16 and 17, the cathode electrode and the anode electrode are all examples made with black lines. For the purpose of distinction, the anode electrode is represented by a dotted line. However, in actual applications, the cathode electrode and the anode electrode correspond to different grayscale levels respectively.

[0238] In an exemplary embodiment, the present application also provides a method for detecting a tail position, as shown in FIG19 , which includes the following steps:

[0239] S1901, controlling the X-ray device to collect the overall X-ray image and corner X-ray image of the target battery cell according to preset device parameters.

[0240] Among them, the equipment parameters of the X-Ray device include current, voltage and amplification power, etc. The equipment parameters of the X-Ray device can be determined by the model information of the target battery cell, such as the cell size and electrode thickness; the overall X-ray image and the corner X-ray image are both 16-bit original format images with depth information; the corner X-ray image includes the four corner areas in the large surface direction of the target battery cell.

[0241] S1902, extracting the region of interest from the overall X-ray image and the corner X-ray image to obtain an overall X-ray image of the region of interest and a corner X-ray image of the region of interest.

[0242] S1903 , performing logarithmic transformation on the overall X-ray image of the region of interest and the corner X-ray image of the region of interest, respectively, to obtain a logarithmic domain overall X-ray image and a logarithmic domain corner X-ray image.

[0243] The logarithmic transformation formula is shown in formula (5). -μd (5)

[0244] S1904 , performing bilateral filtering on the logarithmic domain overall X-ray image and the logarithmic domain corner X-ray image, respectively, to obtain a filtered overall X-ray image and a filtered corner X-ray image.

[0245] S1905 , performing histogram equalization processing on the filtered overall X-ray image and the filtered corner X-ray image respectively to obtain an enhanced overall X-ray image and an enhanced corner X-ray image.

[0246] S1906: Input the overall X-ray image, the enhanced overall X-ray image, and the enhanced corner X-ray image into a position detection model to obtain a tail position detection result of the target battery cell.

[0247] The end position detection result includes normal or abnormal.

[0248] In the embodiment of the present application, X-ray equipment imaging has the characteristics of high penetration and non-invasiveness, and is suitable for bare cell defect detection. The detection method is divided into: X-ray image data collection: obtaining 16-bit original format images with depth information; data preprocessing: improving image quality, eliminating irrelevant information in the image, highlighting image features and enhancing feature detectability through image transformation and correction, extraction of regions of interest, and image enhancement (problems such as low contrast, noise interference, and decreased clarity). The image data quality directly affects the upper limit of the accuracy of the subsequent detection model; position detection model construction and deployment: building an artificial intelligence (AI) algorithm model according to task requirements, considering its application in industrial production, lightweighting the AI ​​algorithm model, and creating an application programming interface (API); training, evaluation and deployment of the position detection model to realize non-destructive detection and identification of the bare cell end position; the position detection model can better meet the real-time monitoring needs of actual industrial production by controlling the frame rate.

[0249] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0250] Based on the same inventive concept, embodiments of the present application also provide a tail position detection device for implementing the aforementioned tail position detection method. The solution provided by this device is similar to the solution described in the aforementioned method. Therefore, the specific limitations of one or more tail position detection device embodiments provided below can be found in the above-described limitations of the tail position detection method and will not be further elaborated here.

[0251] In an exemplary embodiment, as shown in FIG20 , a tail position detection device 2000 is provided, comprising: an image acquisition module 2001 and a result determination module 2002, wherein:

[0252] An image acquisition module 2001 is configured to acquire internal structure images of a target battery cell from multiple different viewing angles in response to a target battery cell end position detection instruction;

[0253] The result determination module 2002 is used to analyze each internal structure image and determine the end position detection result of the target battery cell.

[0254] In one embodiment, the internal structure image includes an overall internal structure image of the target battery cell. As shown in FIG21 , the image acquisition module 2001 includes:

[0255] The first acquisition unit 2101 is used to acquire an overall internal structure image acquired by a first image acquisition device, where the acquisition viewing angle of the first image acquisition device covers the entire surface of the target battery cell.

[0256] In an exemplary embodiment, the internal structure image includes an internal structure image of a corner of a target battery cell. As shown in FIG22 , the image acquisition module 2001 includes:

[0257] The second acquisition unit 2201 is used to acquire a first internal structure image acquired by a second image acquisition device and a second internal structure image acquired by a third image acquisition device; the acquisition angle of the second image acquisition device covers the area where the first corner of the target battery cell is located and the area where the second corner is located; the acquisition angle of the third image acquisition device covers the area where the third corner of the target battery cell is located and the area where the fourth corner is located;

[0258] The third acquiring unit 2202 is configured to merge the first internal structure image and the second internal structure image to obtain a corner internal structure image.

[0259] In an exemplary embodiment, each internal structure image includes an overall internal structure image and a corner internal structure image of the target battery cell; as shown in FIG23 , the result determination module 2002 includes:

[0260] An image enhancement unit 2301 is configured to perform image enhancement processing on the overall internal structure image and the corner internal structure image, respectively, to obtain an enhanced overall internal structure image and an enhanced corner internal structure image;

[0261] The result determination unit 2302 is configured to determine a detection result of the tail position of the target battery cell according to the overall internal structure image, the enhanced overall internal structure image, and the enhanced corner internal structure image.

[0262] In an exemplary embodiment, as shown in FIG24 , the image enhancement unit 2301 includes:

[0263] The first determining subunit 2401 is configured to determine edge pixels of the target image according to the grayscale value of each pixel in the target image; the target image is an overall internal structure image or a corner internal structure image;

[0264] The acquisition subunit 2402 is configured to acquire an image of a region of interest from the target image according to edge pixels of the target image and a preset outward expansion offset;

[0265] The processing subunit 2403 is configured to perform image processing on the region of interest image to obtain an enhanced internal structure image of the target image.

[0266] In an exemplary embodiment, as shown in FIG25 , the processing subunit 2403 includes:

[0267] The conversion subunit 2501 is used to perform logarithmic domain conversion on the image of the region of interest to obtain an internal structure image in the logarithmic domain;

[0268] The filtering subunit 2502 is configured to perform bilateral filtering on the logarithmic domain internal structure image to obtain a filtered internal structure image;

[0269] The equalization subunit 2503 is used to perform equalization processing on the filtered internal structure image to obtain an enhanced internal structure image of the target image.

[0270] In an exemplary embodiment, as shown in FIG26 , the result determination unit 2302 includes:

[0271] The first obtaining subunit 2601 is used to input the overall internal structure image, the enhanced overall internal structure image and the enhanced corner internal structure image into the position detection model to obtain the end position detection result of the target battery cell.

[0272] In an exemplary embodiment, the position detection model includes a depthwise separable convolutional network and an output network; as shown in FIG27 , the first obtaining subunit 2601 includes:

[0273] The second obtaining subunit 2701 is used to perform multi-scale feature extraction on the overall internal structure image, the enhanced overall internal structure image, and the enhanced corner internal structure image through a depthwise separable convolutional network, thereby obtaining overall feature images, enhanced overall feature images, and enhanced corner feature images at multiple different scales.

[0274] The third subunit 2702 is used to input each overall feature image, each enhanced overall feature image and each enhanced corner feature image into the output network to obtain the end position detection result.

[0275] In an exemplary embodiment, as shown in FIG28 , the result determination unit 2302 includes:

[0276] The second determining subunit 2801 is configured to determine a first probability value that the tail position of the target battery cell is normal based on the overall internal structure image and a preset standard overall internal structure image;

[0277] The third determining subunit 2802 is configured to determine a second probability value of whether the tail position of the target battery cell is normal based on the enhanced overall internal structure image and the preset standard enhanced overall internal structure image;

[0278] The fourth determining subunit 2803 is configured to determine a third probability value that the tail position of the target battery cell is normal based on the enhanced corner internal structure image and a preset standard enhanced corner internal structure image;

[0279] The fifth determining subunit 2804 is configured to determine a detection result of the end position of the target battery cell according to the first probability value, the second probability value, and the third probability value.

[0280] In an exemplary embodiment, as shown in FIG29 , the fifth determining subunit 2804 includes:

[0281] The weighting subunit 2901 is configured to obtain a comprehensive probability value based on the first probability value, the second probability value, and the third probability value;

[0282] The sixth determining subunit 2902 is configured to determine that the end position detection result is normal when the comprehensive probability value is greater than a preset probability threshold;

[0283] The seventh determining subunit 2903 is configured to determine that the tail position detection result is abnormal when the comprehensive probability value is less than or equal to the probability threshold.

[0284] Each module in the above-mentioned tail position detection device can be implemented in whole or in part through software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in the computer device in the form of software, so that the processor can call and execute the corresponding operations of each module.

[0285] In one embodiment, a computer device is further provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps in the above method embodiments when executing the computer program.

[0286] The implementation principles and technical effects of each step implemented by the processor in the embodiment of the present application are similar to the principles of the above-mentioned end position detection method, and will not be repeated here.

[0287] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0288] The implementation principles and technical effects of the various steps implemented when the computer program in the embodiment of the present application is executed by the processor are similar to the principles of the above-mentioned end position detection method and will not be repeated here.

[0289] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0290] The implementation principles and technical effects of the various steps implemented when the computer program in the embodiment of the present application is executed by the processor are similar to the principles of the above-mentioned end position detection method and will not be repeated here.

[0291] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0292] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may 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 may include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processor involved in the various embodiments provided herein may be, but are not limited to, a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, and the like.

[0293] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, 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.

[0294] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A method for detecting an end position, wherein, The method includes: In response to a detection instruction for the end position of a target battery cell, obtaining internal structure images of the target battery cell from multiple different perspectives; Analyzing each of the internal structure images to determine the detection result of the end position of the target battery cell.

2. The method according to claim 1, wherein, The internal structure images include the overall internal structure image of the target battery cell. The obtaining of the internal structure images of the target battery cell from multiple different perspectives includes: Obtaining the overall internal structure image collected by a first image acquisition device, and the acquisition perspective of the first image acquisition device covers the entire surface of the target battery cell.

3. The method according to claim 1 or 2, wherein The internal structure images include the corner internal structure image of the target battery cell. The obtaining of the internal structure images of the target battery cell from multiple different perspectives includes: Obtaining a first internal structure image collected by a second image acquisition device and a second internal structure image collected by a third image acquisition device; the acquisition perspective of the second image acquisition device covers the first corner area and the second corner area of the target battery cell; the acquisition perspective of the third image acquisition device covers the third corner area and the fourth corner area of the target battery cell; Combining the first internal structure image and the second internal structure image to obtain the corner internal structure image.

4. The method according to claim 1 or 2, wherein Each of the internal structure images includes the overall internal structure image and the corner internal structure image. The analyzing of each of the internal structure images to determine the detection result of the end position of the target battery cell includes: Performing image enhancement processing on the overall internal structure image and the corner internal structure image respectively to obtain an enhanced overall internal structure image and an enhanced corner internal structure image; Determining the detection result of the end position of the target battery cell according to the overall internal structure image, the enhanced overall internal structure image, and the enhanced corner internal structure image.

5. The method according to claim 4, wherein The performing of image enhancement processing on the overall internal structure image and the corner internal structure image respectively to obtain an enhanced overall internal structure image and an enhanced corner internal structure image includes: Determining the edge pixel points of the target image according to the gray values of each pixel point in the target image; the target image is the overall internal structure image or the corner internal structure image; Obtaining a region of interest image from the target image according to the edge pixel points of the target image and a preset outward expansion offset; Performing image processing on the region of interest image to obtain the enhanced internal structure image of the target image.

6. The method according to claim 5, wherein, The performing of image processing on the region of interest image to obtain the enhanced internal structure image of the target image includes: Converting the region of interest image into the logarithmic domain to obtain a logarithmic domain internal structure image; Performing bilateral filtering processing on the logarithmic domain internal structure image to obtain a filtered internal structure image; Performing equalization processing on the filtered internal structure image to obtain the enhanced internal structure image of the target image.

7. The method according to claim 4, wherein The determining of the detection result of the end position of the target battery cell according to the overall internal structure image, the enhanced overall internal structure image, and the enhanced corner internal structure image includes: Input the overall internal structure image, the enhanced overall internal structure image, and the enhanced corner internal structure image into a position detection model to obtain the detection result of the end position of the target battery cell.

8. The method according to claim 7, wherein The position detection model includes a depthwise separable convolutional network and an output network; the step of inputting the overall internal structure image, the enhanced overall internal structure image, and the enhanced corner internal structure image into the position detection model to obtain the detection result of the end position of the target battery cell includes: Perform multi-scale feature extraction on the overall internal structure image, the enhanced overall internal structure image, and the enhanced corner internal structure image respectively through the depthwise separable convolutional network to obtain overall feature images, enhanced overall feature images, and enhanced corner feature images at multiple different scales respectively; Input each of the overall feature images, each of the enhanced overall feature images, and each of the enhanced corner feature images into the output network to obtain the detection result of the end position.

9. The method according to claim 4, wherein The step of determining the detection result of the end position of the target battery cell according to the overall internal structure image, the enhanced overall internal structure image, and the enhanced corner internal structure image includes: Determine a first probability value that the end position of the target battery cell is normal according to the overall internal structure image and a standard overall internal structure image; determine a second probability value that the end position of the target battery cell is normal according to the enhanced overall internal structure image and a standard enhanced overall internal structure image; determine a third probability value that the end position of the target battery cell is normal according to the enhanced corner internal structure image and a standard enhanced corner internal structure image; Determine the detection result of the end position of the target battery cell according to the first probability value, the second probability value, and the third probability value.

10. The method according to claim 9, wherein, The step of determining the detection result of the end position of the target battery cell according to the first probability value, the second probability value, and the third probability value includes: Determine a comprehensive probability value according to the first probability value, the second probability value, and the third probability value; When the comprehensive probability value is greater than a probability threshold, determine that the detection result of the end position is normal; when the comprehensive probability value is less than or equal to the probability threshold, determine that the detection result of the end position is abnormal.

11. A tail position detection device, wherein, The device includes: An image acquisition module, configured to acquire internal structure images of the target battery cell from multiple different perspectives in response to a detection instruction for the end position of the target battery cell; A result determination module, configured to analyze each of the internal structure images to determine the detection result of the end position of the target battery cell.

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

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

Images