Method, device, system and storage medium for detecting abnormality of battery cell pole piece

Abstract

The present invention discloses a method, device, system and storage medium for detecting abnormalities in the electrode of a battery cell. The detection method comprises the following steps: inputting an electrode image obtained by an X-Ray device into a pre-trained first AI model, and the first AI model determines whether it is OK or NG. If it is determined to be NG, it is output as a definite defective product; receiving an electrode image that is determined to be OK, and determining that any layer's overhang length or angle exceeds the set specifications as a defective product, otherwise it is determined to be OK; receiving an electrode image that is determined to be OK, and determining that the overhang value of a specific layer is less than a set critical value as NO, otherwise it is a good product; receiving and inputting the electrode image that is determined to be NO into a pre-trained second AI model for image similarity matching, and determining that it is NG as a defective product, and determining that it is OK as a critical defective product. The present invention can accurately detect battery cells with abnormal electrode features, reduce misjudgments, and thus improve the overall quality rate of the battery.

Description

Technical Field

[0001] The present invention belongs to the technical field of battery cell detection, and specifically relates to a method, device, system and storage medium for detecting abnormality of a battery cell electrode. Background Art

[0002] With the maturity of image algorithms, AI detection technology has been widely used in the field of X-ray automatic inspection. During X-ray testing of lithium batteries, AI intelligent algorithms are often used to identify and judge images and detect irregular and defective products. Existing soft-pack lithium batteries are made of a wound cathode and anode plates and a separator. The total width of the anode is usually greater than that of the cathode. The length of the anode extending beyond the cathode is called the overhang length. Under the penetration of X-rays, vertical lines representing the different layers of the wound battery cell are displayed.

[0003] Existing detection and judgment of abnormalities in soft-pack lithium battery pole pieces is done by using an X-ray machine to automatically inspect lithium batteries. Typically, the image is first subjected to morphological recognition and judgment. Those that fail the test are output as defective products. Those that pass the test are then evaluated for the overhang length and skew angle (the skew relative to the vertical line of the inner layer) of each layer in the image. Any layer with an overhang length or angle outside the specification range is considered a defective product, while any layer with an overhang length or angle within the specification range is considered a good product. During the image morphological recognition and judgment process, defective products can be divided into two categories: one that is 100% confirmed as defective, i.e., a confirmed defective product; and one with slightly deformed pole pieces that are acceptable to the customer, i.e., a critical defective product. These critical defective products can be output as good products. The main factor affecting the quality rate of lithium battery inspections is critical defective products. The morphological characteristics of critical defective products are similar to those of confirmed defective products, making them difficult for AI models to define. This results in many critical defective products being judged as confirmed defective products, leading to misjudgments (i.e., qualified products being judged as unqualified). Summary of the Invention

[0004] One of the purposes of the present invention is to address the deficiencies in the prior art and provide a method for detecting abnormalities in the electrode of a battery cell. This method can accurately detect battery cells with abnormal electrode features, reduce misjudgments, and thus improve the overall quality rate of the battery.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] A method for detecting abnormality of a battery cell electrode comprises the following steps:

[0007] Step 1: Input the electrode image acquired by the X-Ray device into the pre-trained first AI model for image similarity matching. The first AI model then generates an OK or NG judgment result. Cells judged as NG are output as confirmed defective products.

[0008] Step 2: Receive the electrode image that is judged as OK in step 1, obtain the upper and lower endpoints of each layer of overhang in the electrode image, calculate the length and angle of each layer of overhang, and judge it as NG if the length or angle of any layer of overhang exceeds the set specifications. The battery cell corresponding to the electrode image judged as NG is output as a defective product. The electrode image whose length and angle of each layer of overhang are within the set specifications is judged as OK.

[0009] Step 3: Receive the electrode image that is judged as OK in step 2, and compare the overhang value of the specific layer of the electrode image with the set critical value. If the overhang value of the specific layer of the electrode image is greater than the critical value, it is judged as YES. The cell output of the electrode image judged as YES is good. The electrode image whose overhang value of the specific layer of the electrode image is less than the critical value is judged as NO.

[0010] Step 4: Receive the electrode image judged as NO output in step 3, input the electrode image into the pre-trained second AI model, perform image similarity matching, and the second AI model gives an OK or NG judgment result. The battery cell output corresponding to the electrode image judged as NG is a defective product, and the battery cell output corresponding to the electrode image judged as OK is a critical defective product.

[0011] Furthermore, the specific layer of the pole piece image is the outermost layer of the pole piece image or the second outermost layer of the pole piece image.

[0012] Furthermore, in step 1, after the battery cell is transported by the robot, the angle to be measured of the battery cell is moved to the center of the X-Ray tube, and after being irradiated by X-Ray rays, an image is formed in the intensifier and photographed by the CCD to obtain a pole piece image.

[0013] Furthermore, before inputting the pole piece image into the pre-trained first AI model, the pole piece image is converted into a pole piece image in a grayscale image format and is converted without distortion to obtain a preprocessed pole piece image for input into the first AI model.

[0014] Furthermore, before inputting the electrode image into the pre-trained first AI model, abnormal electrode images and normal electrode images of the battery cell in the training data are collected; the abnormal area of ​​the abnormal electrode image is marked, and the annotation text of the abnormal electrode image is generated according to the coordinates and size of the abnormal area; the abnormal electrode image, the normal electrode image, the annotation text, and the image classification category are input into the target detection model for electrode abnormality prediction training, and after the training is completed, the pre-trained first AI model is obtained;

[0015] Before inputting the pole piece image into the pre-trained second AI model, the outermost and / or second outermost abnormal pole piece images and normal pole piece images of the battery cell in the training data are collected, and the degree of skewness and deformation of the outermost and / or second outermost abnormal pole piece images are smaller than those of the abnormal pole piece images in the training data of the first AI model; the abnormal area of ​​the abnormal pole piece image is marked, and the annotation text of the abnormal pole piece image is generated according to the coordinates and size of the abnormal area; the abnormal pole piece image, normal pole piece image, annotation text, and image classification category are input into the target detection model for pole piece abnormality prediction training, and the pre-trained second AI model is obtained after the training is completed.

[0016] Compared with the prior art, the present invention has at least the following beneficial effects:

[0017] If the electrode image of the present invention is judged as NG by the first AI model, a definite defective product is output. The electrode image types judged as OK are defective and good. Because the electrode of the definite defective product has a large degree of abnormality, the first AI model cannot accurately judge the battery cell with a smaller degree of abnormality, so the next step of judgment is required;

[0018] The length and angle of each layer of overhang of the OK electrode image are compared with the set specifications to further determine the type of the electrode image. The NG electrode image type is determined to be defective, and the OK electrode image type is determined to be defective or good. Among them, the defective products in the OK electrode image type are difficult to determine by comparing the length and angle of each layer of overhang, so the next step of determination is required;

[0019] The OK electrode image is further determined by comparing the overhang value of a specific layer with the critical value. The electrode image type determined as YES is a good product, and the electrode image type determined as NO is a defective product. In order to avoid the defective products in the NO electrode image including critical defective products, the next step of determination is required;

[0020] By running the NO electrode image through the second AI model, critical defective products can be distinguished from defective products. The above four steps can accurately output confirmed defective products, defective products, good products, and critical defective products, reducing misjudgments and thus improving the overall quality rate of the battery.

[0021] The second purpose of the present invention is to provide a battery cell electrode abnormality detection device to address the deficiencies in the prior art. The device can accurately detect battery cells with abnormal electrode characteristics, reduce misjudgments, and thus improve the overall quality rate of the battery.

[0022] In order to achieve the above object, the present invention adopts the following technical solutions:

[0023] A device for detecting abnormality of a battery cell electrode, comprising:

[0024] An acquisition module is configured to acquire an image of a pole piece of a battery cell;

[0025] The first determination module is configured to input the electrode image into a pre-trained first AI model to perform image similarity matching. If the image similarity of the electrode image is low, the battery cell corresponding to the electrode image is determined to be defective. If the image similarity of the electrode image is high, the electrode image is output to the next module;

[0026] a calculation module configured to receive the pole piece image from the first determination module, obtain the upper and lower endpoints of each overhang layer of the pole piece image, and calculate the length and angle of each overhang layer of the pole piece image;

[0027] The second determination module is configured to obtain the length and angle of each layer of overhang in the electrode image. If the length and angle of each layer of overhang in the electrode image exceed the set specifications, the battery cell corresponding to the electrode image is determined to be defective. If the length and angle of each layer of overhang in the electrode image are within the set specifications, the electrode image is output to the next module.

[0028] The third determination module is configured to receive the electrode image from the second determination module, and if the overhang value of a specific layer of the electrode image is greater than a set critical value, determine that the cell corresponding to the electrode image is a good product; if the overhang value of a specific layer of the electrode image is less than the set critical value, output the electrode image to the next module;

[0029] The fourth judgment module is configured to receive the electrode image from the third judgment module, input the electrode image into a pre-trained second AI model, and perform image similarity matching. If the image similarity of the electrode image is low, the battery cell corresponding to the electrode image is judged to be a defective product; if the image similarity of the electrode image is high, the battery cell corresponding to the electrode image is judged to be a critical defective product.

[0030] Furthermore, the battery cell electrode abnormality detection device further includes:

[0031] A conversion module is configured to convert the pole piece image into a pole piece image in a grayscale image format before inputting the pole piece image into a pre-trained first AI model, and perform an undistorted conversion to obtain a pre-processed pole piece image for input into the first AI model;

[0032] The first training module is configured to collect abnormal electrode images and normal electrode images of the battery cell in the training data before inputting the electrode images into the pre-trained first AI model; mark the abnormal areas of the abnormal electrode images, and generate annotation text of the abnormal electrode images according to the coordinates and size of the abnormal areas; input the abnormal electrode images, normal electrode images, annotation text, and image classification categories into the target detection model for electrode abnormality prediction training, and obtain the pre-trained first AI model after the training is completed;

[0033] The second training module is configured to collect the outermost and / or sub-outermost abnormal pole piece images and normal pole piece images of the battery cell in the training data before inputting the pole piece images into the pre-trained second AI model, and the degree of skewness and deformation of the outermost and / or sub-outermost abnormal pole piece images is smaller than that of the abnormal pole piece images in the training data of the first training model; mark the abnormal areas of the abnormal pole piece images, and generate the marked text of the abnormal pole piece images according to the coordinates and size of the abnormal areas; input the abnormal pole piece images, normal pole piece images, marked texts, and image classification categories into the target detection model for pole piece abnormality prediction training, and obtain the pre-trained second AI model after the training is completed.

[0034] Furthermore, the types of abnormal pole piece images in the training data of the first training module include: abnormal pole piece images deformed due to lateral force; abnormal pole piece images deformed due to longitudinal force; abnormal pole piece images with overhang angles greater than 30° for more than three consecutive layers.

[0035] The advantages of the detection device and the above-mentioned method for detecting abnormality of the electrode of the battery cell over the prior art are the same, and will not be repeated here.

[0036] The third purpose of the present invention is to provide a battery cell electrode abnormality detection system to address the deficiencies in the prior art. The system can accurately detect battery cells with abnormal electrode characteristics, reduce misjudgments, and thus improve the overall quality rate of the battery.

[0037] In order to achieve the above object, the present invention adopts the following technical solutions:

[0038] A battery cell electrode abnormality detection system includes an X-Ray device, a processor, and a storage device. The storage device stores a computer program, and the processor is used to execute the program to implement the electrode abnormality detection method of the above embodiment.

[0039] The advantages of the detection system and the above-mentioned method for detecting abnormality of the electrode of the battery cell over the prior art are the same, and will not be repeated here.

[0040] The fourth object of the present invention is to provide a computer-readable storage medium to address the deficiencies in the prior art, which can accurately detect battery cells with abnormal electrode features, reduce misjudgments, and thus improve the overall quality rate of the battery.

[0041] In order to achieve the above object, the present invention adopts the following technical solutions:

[0042] A computer-readable storage medium stores a computer program, which is executed by a processor to implement the pole piece abnormality detection method of the above embodiment.

[0043] The advantages of the storage medium and the above-mentioned method for detecting abnormality of the electrode of the battery cell over the prior art are the same, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Features, advantages, and technical effects of exemplary embodiments of the present invention will be described below with reference to the accompanying drawings.

[0045] Figure 1 This is a diagram explaining the meaning of the pole piece image in an embodiment of the present invention.

[0046] Figure 2 This is a schematic diagram of an electrode image of a confirmed defective product in an embodiment of the present invention.

[0047] Figure 3 Schematic diagram of an electrode image of a critical defective product according to an embodiment of the present invention.

[0048] Figure 4 Flowchart of a method for detecting abnormality of a battery cell electrode according to an embodiment of the present invention.

[0049] Figure 5 Schematic diagram of the structure of the battery cell electrode abnormality detection device in an embodiment of the present invention.

[0050] Figure 6 Schematic diagram of the structure of the battery cell electrode abnormality detection system in an embodiment of the present invention.

[0051] The description of the accompanying drawings is as follows:

[0052] 1-X-Ray device; 2-processor; 3-storage device; 401-acquisition module; 402-conversion module; 403-first training module; 404-first determination module; 405-computing module; 406-second determination module; 407-third determination module; 408-second training module; 409-fourth determination module. DETAILED DESCRIPTION

[0053] For example, certain words are used in the specification and claims to refer to specific components. Those skilled in the art should understand that hardware manufacturers may use different terms to refer to the same component. This specification and claims do not use differences in names as a way to distinguish components, but use differences in the functions of the components as the criteria for distinction. For example, "including" mentioned throughout the specification and claims is an open term, so it should be interpreted as "including but not limited to". "Approximately" means that within an acceptable error range, those skilled in the art can solve technical problems within a certain error range and basically achieve technical effects.

[0054] Furthermore, the terms “first,” “second,” etc. are used for descriptive purposes only and are not to be understood as indicating or implying relative importance.

[0055] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or more executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may be performed out of the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present application belong.

[0056] It should be understood that various parts of the present application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiment, the N steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used to implement the method: a discrete logic circuit having logic gate circuits for implementing logical functions on data signals, an application-specific integrated circuit having suitable combinational logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc. The present invention is further described in detail below with reference to the accompanying drawings, but is not intended to limit the present invention.

[0057] like Figures 1 to 4 As shown, a method for detecting abnormality of a battery cell electrode includes the following steps:

[0058] Step 1: Input the electrode image acquired by the X-Ray device 1 into the pre-trained first AI model to perform image similarity matching. The first AI model then generates an OK or NG judgment result. Cells judged as NG are output as confirmed defective products.

[0059] Step 2: Receive the electrode image that is judged as OK in step 1, obtain the upper and lower endpoints of each layer of overhang in the electrode image, calculate the length and angle of each layer of overhang, and judge it as NG if the length or angle of any layer of overhang exceeds the set specifications. The battery cell corresponding to the electrode image judged as NG is output as a defective product. The electrode image whose length and angle of each layer of overhang are within the set specifications is judged as OK.

[0060] Step 3: Receive the electrode image that is judged as OK in step 2, and compare the overhang value of the specific layer of the electrode image with the set critical value. If the overhang value of the specific layer of the electrode image is greater than the critical value, it is judged as YES. The cell output of the electrode image judged as YES is good. The electrode image whose overhang value of the specific layer of the electrode image is less than the critical value is judged as NO.

[0061] Step 4: Receive the electrode image judged as NO output in step 3, input the electrode image into the pre-trained second AI model, perform image similarity matching, and the second AI model gives an OK or NG judgment result. The battery cell output corresponding to the electrode image judged as NG is a defective product, and the battery cell output corresponding to the electrode image judged as OK is a critical defective product.

[0062] In the above steps 1 and 4, OK means that the electrode lines of each layer are clear and without deflection, and NG means that the electrode lines are skewed, interrupted, or missing. The electrode line is the line segment where the anode electrode piece exceeds the cathode electrode piece in the manually selected electrode piece image.

[0063] Compared with the prior art, the present invention has at least the following beneficial effects:

[0064] If the electrode image of the present invention is judged as NG by the first AI model, a definite defective product is output. The electrode image types judged as OK are defective and good. Because the electrode of the definite defective product has a large degree of abnormality, the first AI model cannot accurately judge the battery cell with a smaller degree of abnormality, so the next step of judgment is required;

[0065] The length and angle of each layer of overhang of the OK electrode image are compared with the set specifications to further determine the type of the electrode image. The NG electrode image type is determined to be defective, and the OK electrode image type is determined to be defective or good. Among them, the defective products in the OK electrode image type are difficult to determine by comparing the length and angle of each layer of overhang, so the next step of determination is required;

[0066] The OK electrode image is further determined by comparing the overhang value of a specific layer with the critical value. The electrode image type determined as YES is a good product, and the electrode image type determined as NO is a defective product. In order to avoid the defective products in the NO electrode image including critical defective products, the next step of determination is required;

[0067] The second AI model verifies NO electrode images, distinguishing critically defective products from defective ones. The four-step process accurately outputs confirmed defectives, defectives, good products, and critically defective products, reducing misjudgments and improving the overall battery quality.

[0068] Furthermore, the specific layer of the pole piece image is the outermost layer of the pole piece image or the second outermost layer of the pole piece image. Usually, the outermost and second outermost pole pieces are easily deformed by external force, so the overhang of the outermost and second outermost layers is more prone to skew than other layers, and the skewness will reduce the overhang value. Therefore, the overhang value of the outermost and second outermost layers can be compared with the set critical value. If it is less than the threshold value, it indicates that there is a risk of skewness, and the pole piece image is output to the second AI model. If it is greater than the threshold value, the outer layer of the surface is not skewed by external force, and the battery cell corresponding to the pole piece image is output as a good product. The critical value for judgment is usually related to the type of battery, and can be 0 to 1 times the overhang value.

[0069] Furthermore, in step 1, after the cell is moved by the robot, the angle to be measured is moved to the center of the X-ray tube. After being irradiated by X-rays, it is imaged in the intensifier and photographed by the CCD to obtain the electrode image. Imaging in the intensifier makes the obtained electrode image clearer, thereby reducing the impact of image noise.

[0070] Furthermore, before inputting the polar slice image into the pre-trained first AI model, the polar slice image is converted into a polar slice image in grayscale image format and subjected to distortion-free conversion to obtain a pre-processed polar slice image for input into the first AI model. By converting the polar slice image into a grayscale image format, the amount of original image data is reduced, facilitating less computational effort in subsequent processing.

[0071] Furthermore, before inputting the electrode image into the pre-trained first AI model, abnormal electrode images and normal electrode images of the battery cell in the training data are collected; the abnormal area of ​​the abnormal electrode image is marked, and the annotation text of the abnormal electrode image is generated according to the coordinates and size of the abnormal area; the abnormal electrode image, the normal electrode image, the annotation text, and the image classification category are input into the target detection model for electrode abnormality prediction training, and after the training is completed, the pre-trained first AI model is obtained;

[0072] Before inputting the pole piece image into the pre-trained second AI model, the outermost and / or second outermost abnormal pole piece images and normal pole piece images of the battery cell in the training data are collected, and the degree of skewness and deformation of the outermost and / or second outermost abnormal pole piece images are smaller than those of the abnormal pole piece images in the training data of the first AI model; the abnormal area of ​​the abnormal pole piece image is marked, and the annotation text of the abnormal pole piece image is generated according to the coordinates and size of the abnormal area; the abnormal pole piece image, normal pole piece image, annotation text, and image classification category are input into the target detection model for pole piece abnormality prediction training, and the pre-trained second AI model is obtained after the training is completed.

[0073] Among them, the abnormal pole piece images in the training data of the first AI model are abnormal pole piece images deformed due to lateral force; abnormal pole piece images deformed due to longitudinal force; abnormal pole piece images with overhang angles greater than 30° for more than three consecutive layers.

[0074] Figure 5 This is a schematic diagram of the structure of the battery cell electrode abnormality detection device provided in an embodiment of the present application. The detection device includes:

[0075] The acquisition module 401 is configured to acquire an image of a pole piece of a battery cell;

[0076] The first determination module 404 is configured to input the electrode image into a pre-trained first AI model to perform image similarity matching. If the image similarity of the electrode image is low, the battery cell corresponding to the electrode image is determined to be defective. If the image similarity of the electrode image is high, the electrode image is output to the next module.

[0077] The calculation module 405 is configured to receive the pole piece image from the first determination module 404, obtain the upper and lower endpoints of each overhang layer of the pole piece image, and calculate the length and angle of each overhang layer of the pole piece image;

[0078] The second determination module 406 is configured to obtain the length and angle of each layer of overhang of the electrode image. If the length and angle of each layer of overhang of the electrode image exceed the set specifications, the battery cell corresponding to the electrode image is determined to be defective. If the length and angle of each layer of overhang of the electrode image are within the set specifications, the electrode image is output to the next module.

[0079] The third determination module 407 is configured to receive the electrode image from the second determination module 406, and if the overhang value of a specific layer of the electrode image is greater than a set critical value, determine that the cell corresponding to the electrode image is a good product; if the overhang value of a specific layer of the electrode image is less than the set critical value, output the electrode image to the next module;

[0080] The fourth determination module 409 is configured to receive the electrode image from the third determination module 407, input the electrode image into a pre-trained second AI model, and perform image similarity matching. If the image similarity of the electrode image is low, the battery cell corresponding to the electrode image is determined to be a defective product; if the image similarity of the electrode image is high, the battery cell corresponding to the electrode image is determined to be a critical defective product.

[0081] Furthermore, the battery cell electrode abnormality detection device further includes:

[0082] The conversion module 402 is configured to convert the polar slice image into a polar slice image in a grayscale image format before inputting the polar slice image into the pre-trained first AI model, and perform an undistorted conversion to obtain a pre-processed polar slice image for input into the first AI model;

[0083] The first training module 403 is configured to collect abnormal electrode images and normal electrode images of the battery cell in the training data before inputting the electrode images into the pre-trained first AI model; mark the abnormal areas of the abnormal electrode images, and generate annotation text of the abnormal electrode images according to the coordinates and size of the abnormal areas; input the abnormal electrode images, normal electrode images, annotation text, and image classification categories into the target detection model to perform electrode abnormality prediction training, and obtain the pre-trained first AI model after the training is completed;

[0084] The second training module 408 is configured to collect the outermost and / or sub-outermost abnormal pole piece images and normal pole piece images of the battery cell in the training data before inputting the pole piece images into the pre-trained second AI model, wherein the degree of skewness and deformation of the outermost and / or sub-outermost abnormal pole piece images is smaller than that of the abnormal pole piece images in the training data of the first training model; mark the abnormal areas of the abnormal pole piece images, and generate the marked text of the abnormal pole piece images according to the coordinates and size of the abnormal areas; input the abnormal pole piece images, normal pole piece images, marked texts, and image classification categories into the target detection model for pole piece abnormality prediction training, and obtain the pre-trained second AI model after the training is completed.

[0085] Furthermore, the types of abnormal pole piece images in the training data of the first training module 403 include: abnormal pole piece images deformed due to lateral force; abnormal pole piece images deformed due to longitudinal force; abnormal pole piece images with overhang angles greater than 30° for more than three consecutive layers.

[0086] Figure 6This is a schematic diagram of the structure of a battery cell electrode anomaly detection system provided in an embodiment of the present application. The detection system may include: an X-ray device 1, a processor 2, and a storage device 3. The storage device 3 stores a computer program, and the processor 2 is used to execute the program. When the processor 2 executes the program, the battery cell electrode anomaly detection method provided in the above embodiment is implemented.

[0087] Furthermore, the detection device also includes a communication interface for communication between the X-Ray device 1 , the storage device 3 and the processor 2 .

[0088] The storage device 3 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0089] If the X-Ray device 1, storage device 3, processor 2, and communication interface are implemented independently, the X-Ray device 1, communication interface, storage device 3, and processor 2 can be connected to each other via a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 6 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0090] Optionally, in a specific implementation, if the storage device 3, processor 2 and communication interface are integrated on a chip, the storage device 3, processor 2 and communication interface can communicate with each other through an internal interface and communicate with the X-Ray device 1 through an external device interconnection bus.

[0091] The processor 2 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0092] This embodiment further provides a computer-readable storage medium on which a computer program is stored. When the program is executed by the processor 2, the above-mentioned method for detecting abnormality of a battery cell electrode is implemented.

[0093] Those skilled in the art will understand that all or part of the steps in the method of the above embodiment can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.

Claims

1. A method for detecting abnormality of a battery cell electrode, characterized in that: The following steps are involved: Step 1: Input the electrode image acquired by the X-Ray device into the pre-trained first AI model for image similarity matching. The first AI model then generates an OK or NG judgment result. Cells judged as NG are output as confirmed defective products. Step 2: Receive the electrode image determined as OK in step 1, obtain the upper and lower endpoints of each layer of overhang in the electrode image, calculate the length and angle of each layer of overhang, and determine that any layer of overhang length or angle exceeds the set specification as NG. The battery cell corresponding to the electrode image determined as NG is output as a defective product. The electrode image whose overhang length and angle are within the set specification is determined as OK. Step 3: Receive the electrode image determined as OK in step 2, and compare the overhang value of the outermost or second outermost layer of the electrode image with a set critical value. If the overhang value of the outermost or second outermost layer of the electrode image is greater than the critical value, it is determined to be YES. The cell corresponding to the electrode image determined to be YES is output as a good product. If the overhang value of the outermost or second outermost layer of the electrode image is less than the critical value, it is determined to be NO. Step 4: Receive the electrode image judged as NO output in step 3, input the electrode image into a pre-trained second AI model, perform image similarity matching, and the second AI model gives a judgment result of OK or NG. The battery cell output corresponding to the electrode image judged as NG is a defective product, and the battery cell output corresponding to the electrode image judged as OK is a critical defective product.

2. The method for detecting anomalies of a pole piece according to claim 1, wherein: In step 1, after the battery cell is transported by a robot, the angle to be measured of the battery cell is moved to the center of the X-Ray tube. After being irradiated by X-Ray rays, the cell is imaged in the intensifier and photographed by a CCD to obtain the electrode image.

3. The method for detecting anomalies of a pole piece according to claim 2, wherein: Before inputting the pole piece image into the pre-trained first AI model, the pole piece image is converted into the pole piece image in grayscale image format and is converted without distortion to obtain the preprocessed pole piece image for input into the first AI model.

4. The method for detecting anomalies of a pole piece according to claim 1, wherein: Before inputting the electrode image into the pre-trained first AI model, collecting abnormal electrode images and normal electrode images of the battery cell in the training data; marking the abnormal area of ​​the abnormal electrode image, and generating the annotation text of the abnormal electrode image according to the coordinates and size of the abnormal area; Inputting the abnormal pole piece image, the normal pole piece image, the annotated text, and the image classification category into the target detection model for pole piece abnormality prediction training, and obtaining the pre-trained first AI model after the training is completed; Before inputting the electrode image into the pre-trained second AI model, collecting abnormal electrode images and normal electrode images of the outermost and / or sub-outer layers of the battery cell in the training data, wherein the degree of skewness and deformation of the abnormal electrode images of the outermost and / or sub-outer layers is smaller than that of the abnormal electrode images in the training data of the first AI model; annotating the abnormal region of the abnormal electrode image, and generating an annotation text for the abnormal electrode image according to the coordinates and size of the abnormal region; The abnormal pole piece image, the normal pole piece image, the annotated text, and the image classification category are input into the target detection model for pole piece abnormality prediction training, and the pre-trained second AI model is obtained after the training is completed.

5. A device for detecting abnormality of a battery cell electrode, characterized in that: include: An acquisition module is configured to acquire an image of a pole piece of a battery cell; A first determination module is configured to input the electrode image into a pre-trained first AI model to perform image similarity matching. If the image similarity of the electrode images is low, the battery cell corresponding to the electrode image is determined to be a definite defective product. If the image similarity of the electrode images is high, the electrode image is output to the next module; a calculation module configured to receive the pole piece image from the first determination module, obtain the upper and lower endpoints of each layer of overhang in the pole piece image, and calculate the length and angle of each layer of overhang in the pole piece image; A second determination module is configured to obtain the length and angle of the overhang of each layer of the electrode image. If the length and angle of the overhang of each layer of the electrode image exceed set specifications, the battery cell corresponding to the electrode image is determined to be defective. If the length and angle of the overhang of each layer of the electrode image are within the set specifications, the electrode image is output to the next module; A third determination module is configured to receive the electrode image from the second determination module, and determine that the cell corresponding to the electrode image is a good product if the overhang value of the outermost layer or the second outermost layer of the electrode image is greater than a set critical value; and output the electrode image to the next module if the overhang value of the outermost layer or the second outermost layer of the electrode image is less than a set critical value; The fourth judgment module is configured to receive the electrode image of the third judgment module, input the electrode image into a pre-trained second AI model, and perform image similarity matching. If the image similarity of the electrode image is high, the battery cell corresponding to the electrode image is determined to be a critical defective product; if the image similarity of the electrode image is low, the battery cell corresponding to the electrode image is determined to be a defective product.

6. The electrode abnormality detection device according to claim 5, characterized in that: Also includes: a conversion module configured to convert the pole piece image into a grayscale image format and perform an undistorted conversion before inputting the pole piece image into the pre-trained first AI model, to obtain the pre-processed pole piece image for input into the first AI model; The first training module is configured to collect abnormal electrode images and normal electrode images of the battery cell in the training data before inputting the electrode images into the pre-trained first AI model; mark the abnormal areas of the abnormal electrode images, and generate annotation text for the abnormal electrode images according to the coordinates and dimensions of the abnormal areas; Inputting the abnormal pole piece image, the normal pole piece image, the annotated text, and the image classification category into the target detection model for pole piece abnormality prediction training, and obtaining the pre-trained first AI model after the training is completed; The second training module is configured to collect, before inputting the electrode image into the pre-trained second AI model, abnormal electrode images and normal electrode images of the outermost and / or second outermost layers of the battery cell in the training data, wherein the degree of skewness and deformation of the abnormal electrode images of the outermost and / or second outermost layers is smaller than that of the abnormal electrode images in the training data of the first training model; annotate the abnormal area of ​​the abnormal electrode image, and generate annotated text for the abnormal electrode image according to the coordinates and size of the abnormal area; The abnormal pole piece image, the normal pole piece image, the annotated text, and the image classification category are input into the target detection model for pole piece abnormality prediction training, and the pre-trained second AI model is obtained after the training is completed.

7. The electrode abnormality detection device according to claim 6, characterized in that: The types of abnormal pole piece images in the training data in the first training module include: abnormal pole piece images deformed due to lateral force; abnormal pole piece images deformed due to longitudinal force; abnormal pole piece images with overhang angles greater than 30° for more than three consecutive layers.

8. A battery cell electrode abnormality detection system, characterized in that: It includes an X-Ray device, a processor and a storage device, the storage device stores a computer program, and the processor is used to execute the program to implement the pole piece abnormality detection method according to any one of claims 1 to 4.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the pole piece abnormality detection method as described in any one of claims 1 to 4.

Images