Slitting defect detection method, device and system for battery pole piece

Abstract

The invention discloses a slitting defect detection method, device and system for a battery pole piece, and relates to the technical field of image processing, the method comprises the following steps: carrying out first processing on a to-be-identified image of the battery pole piece to obtain a first defect feature, the area of the first defect feature being greater than a first area threshold; second processing is carried out on the to-be-recognized image to obtain a second defect feature, the area of the second defect feature is larger than a second area threshold and smaller than the first area threshold, and the second area threshold is smaller than the first area threshold; and determining an abnormal defect feature based on the first defect feature and the second defect feature, and marking the abnormal defect feature. According to the method, the to-be-recognized image is processed in parallel through two different processing modes, the defect features in different size ranges are extracted respectively, accurate detection of the slitting defects of the battery pole piece is achieved, and the detection accuracy and efficiency are improved.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a method, device and system for detecting slitting defects of battery pole pieces. Background Art

[0002] As a core component of the new energy industry, the safety and stability of lithium-ion batteries are directly linked to the quality and safety of end products such as new energy vehicles and energy storage systems. The slitting of battery electrodes (such as copper and aluminum foil) is a crucial step in lithium battery manufacturing. However, the slitting process is prone to defects of various sizes, such as burrs, metal shavings, and powder loss. If these defects are not detected promptly and accurately, they can cause micro-shorts or diaphragm punctures in subsequent production steps or during actual battery use, leading to decreased battery performance and even serious safety accidents.

[0003] At present, although image processing technology is used in existing technologies for automated detection, it can usually only identify defects within a single size range or a single type, and cannot accurately detect and distinguish defect features of different size ranges at the same time. In particular, the ability to distinguish defect features of similar sizes but belonging to different levels is insufficient, which can easily lead to misjudgment or missed detection, affecting the comprehensiveness and accuracy of detection.

[0004] Therefore, there is an urgent need for a battery pole piece cutting defect detection method that can comprehensively and accurately identify defect features of different size ranges at the same time, so as to improve the accuracy, comprehensiveness and efficiency of defect detection and meet the growing quality control needs in the lithium battery manufacturing process.

[0005] The above information disclosed in this Background section is included only for enhancement of understanding of the background of the disclosure and therefore it may contain information that does not form the relevant art that is currently known to a person of ordinary skill in the art. Summary of the Invention

[0006] The present application provides a method, device and system for detecting slitting defects of battery pole pieces, so as to solve the problem that existing methods for detecting slitting defects of battery pole pieces are difficult to accurately detect defects of different size ranges.

[0007] This application adopts the following technical solutions: In a first aspect, the present application provides a method for detecting slitting defects of battery pole pieces, comprising: Performing a first process on the image to be identified of the battery electrode to obtain a first defect feature, where the area of ​​the first defect feature is greater than a first area threshold; Performing a second processing on the image to be identified to obtain a second defect feature, where the area of ​​the second defect feature is greater than a second area threshold and smaller than the first area threshold, and the second area threshold is smaller than the first area threshold; Based on the first defect feature and the second defect feature, an abnormal defect feature is determined and the abnormal defect feature is marked.

[0008] This application achieves efficient and accurate detection of battery pole piece cutting defects by parallel processing of defect features of different size thresholds, avoiding omissions or misjudgments caused by single-scale detection, improving detection reliability and adaptability, and meeting quality control requirements in different production scenarios.

[0009] In combination with the first aspect, in an optional implementation, the first processing is implemented based on an image processing method that does not rely on sample training.

[0010] The first processing is implemented based on an image processing method that does not rely on sample training. There is no need to collect and train a large amount of defect sample data in advance, which can effectively reduce implementation complexity, reduce manpower and resource investment, and improve the deployment efficiency and flexibility of the system.

[0011] In conjunction with the first aspect, in an optional implementation, the first processing includes: After the image to be identified is binarized, a binary image is obtained; Scanning the area to be identified in the binary image and marking it according to the continuity of pixels to obtain at least two connected areas; The area of ​​each connected region is calculated, and a connected region having an area greater than a first area threshold is identified as a first defect feature.

[0012] The first processing does not require sample training and can accurately identify defects through binarization, connected area scanning and area threshold judgment, reducing implementation complexity and cost and improving detection efficiency.

[0013] In combination with the first aspect, in an optional implementation, the second processing is implemented based on a neural network model obtained by training defective samples.

[0014] The second processing is implemented based on a neural network model trained on defect samples. By utilizing the self-learning ability of the model, defect features can be automatically extracted and identified, which can effectively improve the accuracy and generalization ability of defect detection, and is especially suitable for complex scenarios and multi-type defect detection tasks.

[0015] In conjunction with the first aspect, in an optional implementation, the method further includes: Based on the image data in which the second defect feature has been confirmed, the neural network model is incrementally trained to dynamically update the neural network model.

[0016] The neural network model is incrementally trained using the confirmed second defect feature image data to dynamically update the model, enabling it to continuously improve detection accuracy, enhance adaptability to new defects or changing working conditions, and achieve continuous optimization of detection performance.

[0017] In conjunction with the first aspect, in an optional implementation, the method further includes: Classifying the abnormal defect features to obtain a third defect feature corresponding to the first defect type and a fourth defect feature corresponding to the second defect type; When the number of the third defect characteristics detected per unit length of the pole piece exceeds a preset third threshold, a first adjustment signal is sent to the die-cutting parameter adjustment system to adjust the size of the slitting blade gap; When it is detected that the fourth defect feature appears continuously for a preset number of times in the pole piece running direction, a second adjustment signal is sent to the die-cutting parameter adjustment system to adjust the size of the feed amount.

[0018] By classifying abnormal defect characteristics and timely adjusting the slitting knife gap and feed amount according to the defect type and quantity, it is possible to effectively reduce electrode defects and improve production stability and product quality.

[0019] In combination with the first aspect, in an optional implementation, the third defect feature is a feature corresponding to a powder loss defect, and the fourth defect feature is a feature corresponding to a burr defect.

[0020] By classifying and identifying abnormal defect characteristics, the slitting knife gap is adjusted in time when the number of powder loss defect characteristics exceeds the limit, and the feed amount is adjusted in time when burr defect characteristics appear continuously, thereby effectively reducing the occurrence of powder loss and burr defects and ensuring the quality and stability of electrode production.

[0021] In conjunction with the first aspect, in an optional implementation, the method further includes: When a pole piece with abnormal defect characteristics is detected, the control marking device sets a defect mark at the location of the abnormal defect characteristics; Based on the defect marking, the pole pieces with abnormal defect characteristics are eliminated.

[0022] By promptly marking abnormal defect characteristics when they are detected and removing electrodes with abnormal defects based on the defect markings, defective electrodes can be effectively prevented from flowing into subsequent processes, further ensuring product quality and production reliability.

[0023] In conjunction with the first aspect, in an optional implementation, electrodes with abnormal defect characteristics are eliminated based on defect markings, including: Obtain the electrode tape length information in real time, and determine the position of the electrode to be rejected based on the electrode tape length information and the identification information of the defect mark; When the interval between consecutive electrode pieces to be removed is less than a preset length threshold, the consecutive electrode pieces to be removed and the electrode pieces in between are removed together; When the interval between consecutive electrode pieces to be rejected is greater than a preset length threshold, only the electrode pieces to be rejected are rejected.

[0024] This rejection method can effectively avoid equipment loss and reduced production efficiency caused by frequent rejection actions, while ensuring the accuracy and efficiency of the rejection operation, further improving the stability of the production line and product quality.

[0025] In a second aspect, the present application further provides a slitting defect detection control device. The slitting defect detection control device includes various modules for executing the slitting defect detection method in the first aspect or any optional implementation of the first aspect. For example, the slitting defect detection control device includes: A first processing module is configured to perform a first processing on the image to be identified of the battery electrode to obtain a first defect feature, wherein the area of ​​the first defect feature is greater than a first area threshold; a second processing module, configured to perform a second processing on the image to be identified to obtain a second defect feature, wherein the area of ​​the second defect feature is greater than a second area threshold and smaller than the first area threshold, and the second area threshold is smaller than the first area threshold; The abnormal defect marking module is used to determine the abnormal defect feature based on the first defect feature and the second defect feature, and mark the abnormal defect feature.

[0026] For more detailed implementation details of the slitting defect detection control device, please refer to the description of any implementation method in the first aspect above.

[0027] In a third aspect, the present application provides a slitting defect detection control device. The slitting defect detection control device includes a memory and a processor. The memory is used to store a computer program or instruction. When the computer program or instruction is executed by the processor, the method of the first aspect or any possible implementation of the first aspect is implemented.

[0028] In a fourth aspect, the present application provides a slitting defect detection and control system. The slitting defect detection and control system includes: The device described in the second or third aspect above; The image acquisition module is connected to the device and is used to collect images of the battery pole pieces after cutting in real time, obtain the image to be identified and send it to the device.

[0029] In combination with the fourth aspect, in an optional implementation, the image acquisition module includes a liquid lens camera with a dynamic focal length or depth of field adjustment function.

[0030] In a fifth aspect, the present application provides a computer-readable storage medium having a computer program or instructions stored therein, which, when executed by a processor, implements the method of the first aspect or any possible implementation of the first aspect.

[0031] In a sixth aspect, the present application provides a computer program product. The computer program product includes a computer program or instructions, and when the computer program or instructions are executed by a processor, the method in the first aspect or any possible implementation of the first aspect is implemented.

[0032] The beneficial effects of the second to sixth aspects above can be referred to the first aspect or any possible implementation of the first aspect, and will not be described in detail here. Based on the implementations provided in the above aspects, this application can also be further combined to provide more implementations.

[0033] Other advantages, objectives and features of the present application will be reflected in part through the following description, and in part will be understood by those skilled in the art through study and practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] 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 the structures shown in these drawings without paying any creative work.

[0035] Figure 1 This is one of the flow charts of the method for detecting battery electrode cutting defects provided in an embodiment of the present application; Figure 2 This is the second flow chart of the method for detecting battery electrode cutting defects provided in an embodiment of the present application; Figure 3 This is the third flow chart of the method for detecting battery electrode cutting defects provided in an embodiment of the present application; Figure 4 This is the fourth flow chart of the method for detecting battery electrode cutting defects provided in an embodiment of the present application; Figure 5 This is one of the structural schematic diagrams of the battery electrode cutting defect detection control device provided in the embodiment of the present application; Figure 6 This is the second structural diagram of the battery electrode cutting defect detection control device provided in an embodiment of the present application; Figure 7 It is a structural diagram of a battery electrode cutting defect detection control system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0036] It should be noted that, in this application, words such as "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described in this application as "exemplary" or "for example" should not be construed as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0037] The term "and / or" as used in this application includes any and all combinations of one or more related listed items. Terms containing ordinal numbers such as "first" and "second" used in this application can be used to describe various constituent elements, but these constituent elements are not limited by these terms. The purpose of using these terms is only to distinguish one constituent element from other constituent elements and cannot be understood as indicating or implying relative importance. For example, without departing from the scope of the rights of this application, the first constituent element can be named as the second constituent element, and similarly, the second constituent element can also be named as the first constituent element.

[0038] Before introducing the embodiments of the present application, the background technology involved in the present application is first introduced.

[0039] The slitting process of lithium battery pole pieces is prone to defects such as burrs, metal shavings, and powder loss, which seriously affect battery safety performance. Traditional detection methods currently suffer from high costs, insufficient precision, and low intelligence, making them unable to accurately detect defects of different types and sizes. Therefore, there is an urgent need to develop an efficient, precise, and intelligent pole piece defect detection method to improve detection accuracy, reduce missed detection rates, and ensure the quality and safety of lithium battery products.

[0040] In summary, the slitting defect detection methods in the related art have the problem of difficulty in accurately detecting defects of different size ranges. The following describes the technical solutions of this application through multiple embodiments. It should be noted that these embodiments can be implemented in a variety of different forms and should not be construed as being limited to the embodiments described herein.

[0041] See Figure 1 , Figure 1It is one of the flow charts of the method for detecting the slitting defects of battery pole pieces provided in the embodiments of the present application. It should be noted that this specification provides the method operation steps as described in the embodiments or flow charts, but more or fewer operation steps may be included based on conventional or non-creative labor. The order of steps listed in the embodiments is only one way of executing the steps among many, and does not represent the only execution order. In practice, when the method program is executed, it can be executed in sequence or in parallel according to the methods shown in the embodiments or the accompanying drawings (for example, in a parallel processor or multi-threaded processing environment). The method can be executed by a controller (such as a PLC or IPC), or by various automated and intelligent devices or systems such as an industrial Internet platform (cloud system) or an AI edge computing device. The embodiments of the present application are not limited to this. In the following embodiments, the method for detecting the slitting defects of battery pole pieces executed by a controller is used as an example for illustrative explanation.

[0042] like Figure 1 As shown, the method for detecting the cutting defects of the battery electrode sheet includes at least the following steps: S101 performs a first process on the image to be identified of the battery electrode to obtain a first defect feature, where the area of ​​the first defect feature is greater than a first area threshold.

[0043] The image to be identified refers to the image captured at the electrode cutting point during the process of detecting the cutting defects of the battery electrode.

[0044] The first processing usually refers to a fast image preprocessing and preliminary defect size classification method, the purpose of which is to quickly screen out the image features of obviously larger defects (area greater than the first area threshold, that is, size greater than 20μm), namely the first defect features.

[0045] It's worth noting that the range of the first area threshold depends on the size range. This size can be, but is not limited to, a key geometric measurement of the defect (such as diameter, length, width, or maximum dimension). The first area threshold can be, for example, the area of ​​a circle with that size as the diameter or the area of ​​a rectangle with that size as the diagonal. The size range can be, but is not limited to, 15μm to 20μm, and can be flexibly set based on actual process requirements. For example, if the size is set to 20μm, and after the electrode is cut, a camera captures an image of the electrode edge to be identified. After the first processing step, this image successfully identifies a distinct burr feature with a size of 25μm, exceeding the set size of 20μm. This burr feature is referred to as the "first defect feature." Simultaneously, a smaller burr, measuring only 8μm, is present in the image, falling below the 20μm threshold and therefore not yet identified and classified by the first processing step.

[0046] In some embodiments, the first processing is implemented based on an image processing method that does not rely on sample training, which can be, but is not limited to, the following methods: 1) Connected region method: after binarization processing of the image, each independent connected region is quickly extracted, and the size information of each region such as area, length, width, etc. is counted. For example, after threshold binarization processing of the collected pole piece edge image, the connected region method is used to quickly mark the large-size burrs with an area exceeding the threshold of 20 μm.

[0047] 2) Image threshold segmentation method: by setting a gray or color threshold, the pole piece image is quickly segmented into background and defect target region, and the region where the defect may exist is preliminarily extracted. For example, by using a fixed threshold or an adaptive threshold method, the obvious defect region (such as metal chips, burrs, and powder loss) in the pole piece image is quickly separated out.

[0048] 3) Edge detection method: classic edge detection operators such as Sobel, Canny, and Prewitt are used to quickly extract the edge features of the pole piece and identify obvious abnormal edges or burr regions. For example, the Canny operator is used to quickly mark the obvious abnormal burr region of the pole piece cutting edge.

[0049] 4) Contour detection method: the outer contour of the target object in the image is quickly extracted, and the size, length, area, and other features of the contour are calculated to screen out large-size abnormal defect features. For example, the findContours() function in the OpenCV library is used to quickly extract the contour of the pole piece defect region for size screening.

[0050] 5) Image projection analysis method: that is, the gray scale change of the image is quickly analyzed along a certain direction (such as horizontal or vertical projection) to preliminarily judge the obvious abnormal burr region.

[0051] For example, the gray scale projection of the pole piece edge image along the horizontal direction is performed to quickly detect the position of the obvious burr.

[0052] S103: The second processing is performed on the to-be-identified image to obtain a second defect feature, the area of the second defect feature is greater than a second area threshold and less than a first area threshold, and the second area threshold is less than the first area threshold.

[0053] Compared with the first processing, the second processing can determine defects with smaller sizes, aiming to improve the detection accuracy.

[0054] Specifically, during the battery electrode slitting defect detection process, while the image to be identified undergoes the first processing, the image can also undergo the second processing in parallel. The second processing is used to identify defect features with an area between a larger threshold (the first area threshold) and a smaller threshold (the second area threshold), namely small and medium-sized defects. The two different image processing methods (i.e., the first processing and the second processing) each identify and analyze defect features within different size ranges (i.e., area ranges). By performing these two processing methods simultaneously (in parallel), large defects can be effectively detected while also accurately identifying small and medium-sized defects, significantly improving detection accuracy and efficiency. It is worth noting that the size corresponding to the second area threshold can range from, but is not limited to, 5μm to 15μm and can be flexibly set based on actual process requirements.

[0055] S105: Determine an abnormal defect feature based on the first defect feature and the second defect feature, and mark the abnormal defect feature.

[0056] Specifically, the aforementioned steps (e.g., S101 and S103) yield defect signatures for two different size ranges: the first defect signature is large defects (with an area greater than a first area threshold) detected quickly through the first process; the second defect signature is small and medium-sized defects (with an area between the first and second area thresholds) detected with high precision through the second process. This step (S105), building on these two defect types, conducts a comprehensive analysis of them to identify abnormal defect signatures (e.g., defects exceeding size specifications or with an abnormal number or frequency), and labels these abnormal defects to provide a basis for subsequent identification, tracing, interception, or rejection.

[0057] For example, the detection backend compares the image feature library trained by artificial intelligence (including defect morphology, color difference, size, position deviation, etc.) with the standard, conducts a comprehensive analysis of the first and second defect features identified previously, and identifies the abnormal defect features that really affect product quality. The identified abnormal defect features can then be marked on the electrode material or edge using various methods such as laser printing and engraving, electronic ink printing or laser punching.

[0058] Assume that during a certain electrode slitting production process, the following defect characteristics are obtained through two parallel detection methods (i.e., S101 and S103): First defect feature (first processing): Two larger defects were detected, namely: There is a burr with a size of 22μm on the edge of the pole piece; there is a granular metal chip with a size of 30μm on the surface of the pole piece.

[0059] The second defect feature (second processing): Three smaller defects were detected, namely: a small burr with a size of 12μm on the edge of the pole piece; and two small particles of powder with sizes of 8μm and 9μm on the surface of the pole piece.

[0060] Next, execute step S105: (1) Determine the characteristics of abnormal defects: The image algorithm software is called to compare and analyze with the AI ​​defect feature library to obtain the following: 22μm burrs and 30μm metal particles: These are large in size and significantly exceed the preset threshold (e.g., 20μm). They are considered significant abnormal defects and should be marked and removed. 12μm burrs, 8μm, and 9μm powders: Although smaller in size, according to AI algorithm analysis, the 12μm burr is located at a critical position on the electrode, which may pose a threat to the diaphragm and is therefore identified as an abnormal defect. The other two powder particles are small in size, but the AI ​​algorithm determines that they will not affect battery performance and are not marked as abnormal for the time being.

[0061] (2) Marking abnormal defect features: The identified abnormal defect locations (22μm burrs, 30μm particles, 12μm burrs) are marked at the corresponding positions of the defective electrodes by laser engraving or electronic inkjet, so that subsequent processes such as CCD cameras can recognize the defect marks and then cut and remove them.

[0062] In general, step S105 is an important link in achieving precise defect control in the technical solution of this application. It accurately determines abnormal defects and implements automatic marking through comprehensive rapid detection and high-precision detection of defect characteristics, thereby realizing intelligent monitoring, precise tracing and effective interception of the electrode slitting process, ensuring high quality and high reliability of battery production, and effectively solving the problem that existing technologies are difficult to achieve efficient and accurate online defect identification, marking and control.

[0063] In addition, the application can monitor the running state of the equipment and software in real time by detecting that the camera shooting and the image algorithm software are logically linked and run in parallel. When the camera communication is abnormal, the display interface immediately alarms. If the image photographed by the camera appears batch distortion, the distortion degree of the photo is judged in real time, and it is compared with the set threshold value, wherein the set threshold value range is 0-100%, and the preferred range is 0.01%-1%. When the image distortion degree exceeds the set threshold value, the display interface immediately displays abnormal information for manual intervention in time. At the same time, if the camera shooting is normal and the image algorithm software is abnormal in online running, real-time alarm is also given, and the photographed photos are transferred to the offline detection module for offline detection analysis (such as manual detection), so that the abnormal products of the electrode sheet cannot flow out with the production process, and the control ability of the defects is improved, and the stability and reliability of the battery production quality are further ensured.

[0064] Referring to Figure 2 , Figure 2 is a flowchart of a battery electrode sheet slitting defect detection method provided by the embodiment of the application. As shown in Figure 2 , in some embodiments, step S101 includes the following steps: S201: after the binaryzation processing of the to-be-identified image, a binary image is obtained; S203: the to-be-identified region in the binary image is scanned, and is marked according to the continuity of the pixels, to obtain at least two connected regions; S205: the area of each connected region is calculated, and the connected region with an area greater than a first area threshold value is identified as a first defect feature.

[0065] Specifically, first, the to-be-identified region (for example, the region of the electrode sheet edge or the slitting part) is extracted from the electrode sheet image photographed by the camera. Then, the original gray-scale image is converted into a binary image with only black and white pixel values (0 and 255) through an image processing method (for example, a gray-scale threshold method). The purpose is to distinguish the background from the defect region, so as to facilitate subsequent processing.

[0066] Then, the obtained binary image is scanned (generally scanned line by line), and different regions are distinguished according to the continuity relationship between the pixels. If a pixel and its adjacent pixels are all the same value (for example, all white pixels 255) in the binary image, these continuous pixels are marked as the same connected region. After scanning and marking, usually a plurality of connected regions are obtained, and each connected region represents a possible defect or background region.

[0067] Next, the area of ​​each marked connected region (i.e., the number of pixels or actual size within the region) is calculated. The calculated area of ​​each connected region is compared with a first area threshold, which is typically determined by actual production requirements (e.g., an area threshold corresponding to a defect size of 20μm). Connected regions whose areas exceed the threshold are preliminarily identified as obvious defects and are referred to as "first defect features."

[0068] Regarding the setting of the first area threshold, for example, assume that an image of the edge area of ​​a battery electrode after slicing is captured. The image size is 1000×1000 pixels, and the pixel resolution is 1 pixel corresponding to 2μm (assuming a scale). According to production requirements, defects larger than 20μm are defined as large defects, namely the "first defect feature." The corresponding area threshold is calculated as follows: The number of pixels corresponding to a scale of 20μm is: 20μm ÷ 2μm / pixel = 10 pixels The area threshold (taking a circular defect as an example) is: π × (radius)^2 = 3.14 × (5 pixels)^2 ≈ 78.5 pixels. 79 pixels is taken as the area threshold.

[0069] Through the above steps, obvious large-size defect features can be preliminarily screened out from the original image (i.e., the image to be identified) quickly and effectively, thereby improving the efficiency of electrode defect detection.

[0070] In some embodiments, the second processing is implemented based on a neural network model trained with defect samples, that is, step S103 may include step S207: recognizing the image to be recognized based on the neural network model trained with defect samples to obtain a second defect feature.

[0071] Specifically, the neural network model is built by pre-collecting a large number of defect image samples (e.g., images of defective areas on battery electrodes) and annotating these samples (e.g., marking defect locations and categories). This annotated defect sample data is then used to train a neural network model (e.g., a convolutional neural network (CNN)) capable of automatically identifying complex or subtle defect features within the images. The trained neural network model is then used to identify newly acquired images on the production line in real time, yielding more refined and complex defect information known as "second defect features."

[0072] After long-term use in actual production environments, neural network models may encounter new defect types or different defect manifestations. To maintain the model's high recognition accuracy, continuous and dynamic model optimization is generally required. To improve the accuracy of the neural network model, in some embodiments, the battery electrode slitting defect detection method further includes: incrementally training the neural network model based on image data in which the second defect characteristic has been confirmed, so as to dynamically update the neural network model.

[0073] The specific method is: after the model identifies a suspected defect area, it is confirmed as a defect again through manual or other means (that is, after confirming the second defect feature), then these newly confirmed defect image data can be added to the original training data set, and the existing neural network model can be incrementally trained again.

[0074] The above steps can identify more complex and subtle defect features (second defect features), making up for the lack of precision or complexity in the first processing step (such as the connected region method). Through incremental learning, the model is continuously and dynamically optimized and updated, gradually improving the model's generalization ability and adaptability, so that a high defect recognition accuracy can be maintained for a long time.

[0075] See Figure 3 , Figure 3 This is the third flow chart of the method for detecting the cutting defects of the battery electrode provided in the embodiment of the present application. Figure 3 As shown, in some embodiments, the method further includes: S301: Classifying abnormal defect features to obtain a third defect feature (i.e., a powder loss defect feature) corresponding to the first defect type and a fourth defect feature (i.e., a large-size burr defect feature) corresponding to the second defect type; S303: When the number of powder loss defect features detected per unit length of the electrode exceeds a preset third threshold, a first adjustment signal is sent to the die-cutting parameter adjustment system to adjust the size of the slitting knife gap; when a large-size burr defect feature is detected to appear continuously for a preset number of times in the electrode tape direction, a second adjustment signal is sent to the die-cutting parameter adjustment system to adjust the size of the feed amount.

[0076] Specifically, in the previous step, some abnormal defect features have been detected by image processing and neural network model, which can be diverse, and different types of defects often correspond to different equipment parameters or process problems. Therefore, the purpose of step S301 is to further classify the detected abnormal defect features, determine the defect categories corresponding to different defect features, and correspond them to specific process problems. For example, the first defect category can be burr defects caused by improper cutting knife gap (denoted as the third defect feature), and the second defect category can be curling or powder dropping defects caused by abnormal feed amount (denoted as the fourth defect feature).

[0077] It is worth noting that the classification of abnormal defect features is usually based on image feature analysis, signal processing or machine learning model, and the specific implementation methods can include but are not limited to the following four methods: 1) Feature analysis method based on image processing: using image features such as defect shape, size, gray level, texture, edge morphology for classification, such as contour analysis, edge detection, area calculation, gray threshold segmentation, etc. For example, according to the visual difference between powder dropping defects (such as uneven surface gray level, spot-like features) and burr defects (edge protrusion, abnormal length-width ratio), the corresponding image features are extracted and thresholded.

[0078] 2) Statistical feature classification method: extract the size, area, aspect ratio, area ratio, gray mean, gray variance, etc. of the defect area as features, and use rule threshold or statistical method (such as cluster analysis) for classification.

[0079] 3) Machine learning classification method: use supervised learning classifier (such as support vector machine SVM, random forest, decision tree, neural network CNN) for training, through labeling sample data of different defect categories, automatically learn feature patterns, and accurately distinguish powder dropping defects and large size burr defects.

[0080] 4) Deep learning image recognition method: use convolutional neural network (CNN) or transfer learning method to train the model with defect image dataset, realize automatic feature extraction and classification, with higher recognition accuracy and generalization ability.

[0081] In actual implementation, one of the above methods can be used alone, or image processing and machine learning methods can be combined to form a hybrid classification model to improve the accuracy, stability and automation of classification.

[0082] Secondly, after classifying the defect characteristics, the frequency or continuity of the defects can be statistically analyzed to determine whether there are any anomalies in the current production process. If the number of third defect characteristics (such as burr defects) appearing per unit length of the electrode (or per unit time) exceeds a certain set threshold (i.e., the third threshold), it means that there may be continuous large-scale burr defects, indicating that the current tool gap may be too large or too small, and the gap needs to be adjusted in a timely manner to reduce the occurrence of defects.

[0083] It should be noted that in a roll-to-roll continuous production model, where electrodes are produced in the form of continuous rolls, the use of "unit length" can clearly indicate whether batch defects occur within a specific length range of the continuous material produced. For example, using "unit electrode length (such as every continuous 1 meter or 10 meters)" as the statistical unit, if powder loss or burr abnormalities are frequently or continuously detected within this unit length range (such as continuous 1 meter or 10 meters), it can be determined to be a batch defect.

[0084] For example, if the algorithm software detects bulk powder loss or continuous large burr defects within a unit electrode length (e.g., 1 or 10 meters), it will transmit a signal to the die-cutting parameter adjustment system, which adjusts the slitting blade gap by 0.1% to 10%. The die-cutting parameter adjustment system specifically includes an electric or pneumatic actuator for adjusting the slitting blade gap size and feed rate, as well as a corresponding control module. If no improvement is seen after adjustment, an alarm will be prompted for manual intervention, and the tool and other abnormalities will be checked.

[0085] At the same time, if the fourth defect feature (such as curling or powder loss defects) appears in the tape-feeding direction of the electrode for multiple consecutive times (i.e., reaching the preset number of times), it means that batch powder loss may occur, indicating that the equipment feed amount (cutting depth) may not be appropriate and the feed amount needs to be adjusted in time.

[0086] By detecting the number or continuity of different defect characteristics, the corresponding adjustment signal is automatically sent to the parameter adjustment system (the first signal adjusts the tool gap, and the second signal adjusts the tool feed amount), realizing automatic feedback and optimization of process parameters.

[0087] See Figure 4 , Figure 4 This is the fourth flow chart of the method for detecting the cutting defects of the battery electrode provided in the embodiment of the present application. Figure 4 As shown, in some embodiments, the method further includes: S401: When a pole piece with an abnormal defect feature is detected, controlling the marking device to set a defect mark at the location of the abnormal defect feature; Based on the defect marking, the pole pieces with abnormal defect characteristics are eliminated.

[0088] Specifically, in the foregoing steps (such as S101-S301), abnormal defect features on the battery pole piece have been identified by image processing algorithms and neural network models. These defects can include burrs, cracks, powder loss, curling, etc., and have been explicitly classified. When an abnormal defect is detected on a certain pole piece, a clear mark needs to be made at the defect location (or the edge thereof) of the pole piece. This mark can be physical (such as inkjet marking, laser marking), or an electronic tag or location record. In subsequent production processing links (such as pole piece winding, stacking, and cell assembly processes), the defective pole pieces can be automatically identified according to the defect marks and removed from the production line, preventing defective pole pieces from entering the next process and ensuring overall product quality.

[0089] For example, if a battery factory pole piece production line produces 100 pole pieces per minute, and the real-time image detection system detects an abnormal defect (such as severe edge powder loss) on the 28th pole piece. At this time, the automatic inkjet marking device immediately marks a red dot on the edge of the 28th pole piece. In the subsequent process, the vision recognition system detects the red dot mark, and the automatic rejection device (such as a blowing nozzle or a mechanical arm) removes the 28th pole piece from the production line. Thus, it is ensured that all pole pieces entering the cell assembly process are qualified products, and the product quality is stable.

[0090] The above steps realize the automatic marking and rejection process of the detected defective pole pieces, effectively preventing defective products from entering subsequent production links, and ensuring the efficiency and reliability of production line quality control.

[0091] In some embodiments, the rejection of the pole piece with abnormal defect features based on the defect mark includes: S403: Real-time acquisition of the pole piece running length information, and determination of the position of the pole piece to be rejected according to the pole piece running length information and the identification information of the defect mark; S405: When the interval between consecutive pole pieces to be rejected is less than a preset length threshold, the consecutive pole pieces to be rejected and the pole pieces between them are rejected together; when the interval between consecutive pole pieces to be rejected is greater than the preset length threshold, only the pole pieces to be rejected are rejected.

[0092] Specifically, on the production line, the pole piece continuously passes through the production equipment at a certain speed. By setting a length sensor (such as a rotary encoder, a photoelectric sensor, or a vision measurement device set on the pole piece transmission roller), the length or position of the pole piece running is measured and recorded in real time. In the previous step (S401), when the pole piece detects an abnormal defect, the device has already marked it (for example, inkjet marking or laser marking). Here, through the vision system or other sensors, the position and information of these defect marks are identified in real time. Based on the real-time measurement of the pole piece position (for example, the length passed from the starting point) and the identification information of the defect marks, it is accurately determined which pole pieces need to be rejected, thereby providing accurate position basis for subsequent automatic rejection action.

[0093] It should be noted that the interval between the continuous to-be-rejected pole pieces refers to the distance or the number of pole pieces between two or more pole pieces marked as "to be rejected" when they appear adjacent to each other on the production line. In order to improve production efficiency and quality control, the production line usually sets a length threshold (for example, 0.5 meters or the length of several pole pieces). If the interval between the to-be-rejected pole pieces is less than this threshold, it means that the defects are concentrated in a certain section of the pole piece. At this time, in order to ensure quality and simplify the rejection action, the intermediate unmarked pole pieces are also rejected together. If the interval between the to-be-rejected pole pieces is greater than this threshold, it means that the defects are scattered and there is no obvious concentration area, and only the individually marked pole pieces need to be rejected.

[0094] For example, if a certain battery pole piece production line continuously runs at a certain speed (for example, 10 meters / minute), real-time detection of slitting defects and marking of pole pieces with defects is performed: Case ①: When the defects are relatively concentrated, assume that the 100th pole piece is marked as having a defect, the 101st pole piece is unmarked, the 102nd pole piece is marked as having a defect, the 103rd pole piece is unmarked, and the 104th pole piece is marked as having a defect. At this time, the interval between the adjacent marked pole pieces (100th, 102nd, and 104th) is very small (only one unmarked pole piece, and the length corresponding to this interval is 0.2 meters), and the preset length threshold is 0.5 meters. Therefore, the system automatically rejects the 100th to 104th pole pieces (including the unmarked 101st and 103rd pole pieces) together to ensure quality stability and avoid missing potential defects.

[0095] Case ②: When the defects are scattered, assume that among the subsequently produced pole pieces, the 200th pole piece is marked as having a defect, and the 210th pole piece is marked as having a defect. At this time, the two marked pole pieces are separated by 9 pole pieces, and the interval corresponding to the length is 1.8 meters, which exceeds the threshold of 0.5 meters. Therefore, the system separately rejects the 200th and 210th pole pieces, without affecting the intermediate unmarked pole pieces.

[0096] As can be seen, step S403 accurately determines the location of the electrode pieces to be rejected through real-time length measurement and mark recognition; step S405 flexibly determines the rejection range based on the comparison of the spacing between the electrode pieces to be rejected with the preset threshold. The overall process ensures a balance between production efficiency and product quality, achieving an intelligent, automated and efficient production process.

[0097] Based on the same technical concept, the present application also provides a battery electrode cutting defect detection control device, referring to Figure 5 , Figure 5 This is one of the structural diagrams of the battery electrode cutting defect detection control device provided in the embodiment of the present application. Figure 5 As shown, the apparatus includes a first processing module 501, a second processing module 502, and an abnormal defect marking module 503, wherein: The first processing module 501 is used to quickly process the collected image to be identified, detect and extract defect features (such as obvious debris, large burrs, etc.) with a large size (i.e., an area larger than a first area threshold); The second processing module 502 is configured to use another more detailed or precise method for the same image to be identified, detecting and extracting defect features that are smaller in size but still have an impact on quality (the area is between the second area threshold and the first area threshold, where the second area threshold is less than the first area threshold). This method is performed in parallel with the first processing module. For example, a deep learning algorithm is used to accurately identify small defects that may affect product performance (such as tiny holes and tiny edge burrs). The abnormal defect marking module 503 is used to comprehensively judge the abnormal defects that actually affect the product quality based on the defect features detected by the first processing module and the second processing module, and mark these abnormal defects for subsequent processing or removal. For example, the detected obvious defects and defects in smaller sensitive areas are uniformly judged as abnormal defects, and marks are made at the corresponding positions to facilitate subsequent processing (such as automatic cutting, removal, etc.).

[0098] In short, the three modules are responsible for quickly screening obvious defects (the first processing module), accurately identifying smaller defects (the second processing module), and comprehensively judging and marking abnormal defects (the abnormal defect marking module).

[0099] Based on the same technical concept, the present application embodiment also provides a slitting defect detection control device, referring to Figure 6 , Figure 6 This is the second structural diagram of the battery electrode cutting defect detection control device provided in the embodiment of the present application. Figure 6 As shown, the device includes a memory 601 and a processor 602. The memory 601 is used to store computer instructions. When the processor 602 executes the computer instructions, the method steps in any method embodiment are implemented.

[0100] The memory 601 includes at least one type of computer-readable storage medium, including flash memory, hard disk, multimedia card, random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), magnetic disk, optical disk, etc. In some embodiments, the computer-readable storage medium may be an internal storage unit of an electronic device, such as a hard disk or memory of the electronic device. In other embodiments, the computer-readable storage medium may also be an external storage device of the electronic device, such as a plug-in hard disk, a secure digital (SD) card, a flash memory card, etc. equipped on the electronic device. Of course, the computer-readable storage medium may also include both an internal storage unit of the electronic device and its external storage device. In this embodiment, the computer-readable storage medium is generally used to store an operating system and various application software installed on the electronic device, such as the program code of the data processing method in the embodiment. In addition, the computer-readable storage medium may also be used to temporarily store various types of data that have been output or are to be output.

[0101] In some embodiments, the processor 602 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other chips. The processor 602 is typically used to control the overall operation of the processing device, such as performing control and processing related to data exchange or communication with other entities. In this embodiment, the processor 602 is used to execute program code stored in the memory 601 or process data.

[0102] Based on the same technical concept, the present application embodiment also provides a slitting defect detection control system, referring to Figure 7 , Figure 7 This is a schematic diagram of the structure of the battery electrode cutting defect detection control system provided in the embodiment of the present application. Figure 7 As shown, the system includes an image acquisition module 701, a slitting defect detection control device 702 and a pole piece cutting module 703, wherein the slitting defect detection control device 702 is a device in any of the above device embodiments.

[0103] Specifically, the slitting defect detection control device 702 is connected to the image acquisition module 701 and the electrode cutting module 703. The image acquisition module 701 is responsible for real-time acquisition of high-definition image data after electrode slitting; the slitting defect detection control device 702 is used to analyze image data in real time, identify and mark abnormal defects; and the electrode cutting module 703 is responsible for automatically cutting and removing abnormal electrodes based on the marked information, forming a closed-loop control system. Through the synergistic effect of these three modules, intelligent defect detection and control are achieved for the battery electrode slitting production line, greatly improving production quality, efficiency, and reliability.

[0104] In some embodiments, the image acquisition module includes a liquid lens camera with dynamic focus or depth of field adjustment function to achieve micron-level image pixel capture, thereby ensuring clear pole piece photography.

[0105] Specifically, the camera uses a new type of liquid lens, which can dynamically adjust the driving voltage at the lens end according to the tiny fluctuations on the surface of the electrode to be detected in real time, and then quickly change the focal length or depth of field of the lens to achieve high-precision image pixel capture at the micron level, meeting the high-resolution requirements of detection size ≤6μm and single pixel <2~3μm, ensuring that the image of the electrode is clear and rich in details.

[0106] In addition, by rationally arranging two groups of liquid lens cameras, images of the pole piece cutting point can be captured simultaneously from both the horizontal and vertical planes of the pole piece, thereby achieving all-round visual inspection of the pole piece edge and effectively identifying various types of minor defect features that may exist during the pole piece cutting process.

[0107] Based on the same technical concept, the embodiment of the present application also provides a computer-readable storage medium, which includes a computer program or instruction stored in the storage medium, and when the computer program or instruction is executed by the processing device, the method steps in any method embodiment are implemented. For more information, please refer to the method embodiment and will not be repeated here. In this embodiment, the computer-readable storage medium includes a flash memory, a hard disk, a multimedia card, a random access memory (RAM), a static random access memory (SRAM), a read-only memory (ROM), a magnetic disk, an optical disk, etc. In some embodiments, the computer-readable storage medium can be an internal storage unit of an electronic device, such as a hard disk or memory of the electronic device. In other embodiments, the computer-readable storage medium can also be an external storage device of an electronic device, such as a plug-in hard disk, a secure digital (SD) card, a flash memory card, etc. equipped on the electronic device. Of course, the computer-readable storage medium can also include both the internal storage unit of the electronic device and its external storage device. In this embodiment, the computer-readable storage medium is generally used to store the operating system and various application software installed on the electronic device, such as the program code of the data processing method in the embodiment. In addition, the computer-readable storage medium can also be used to temporarily store various types of data that have been output or are about to be output.

[0108] Based on the same technical concept, embodiments of the present application further provide a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the data processing method provided in the above method embodiment.

[0109] The above description involves various modules and units. It should be noted that the description of various modules and units is divided into these modules and units for the purpose of clarity. However, in actual implementation, the boundaries of various modules and units may be vague. For example, any or all functional modules and units in this application may share various hardware and / or software elements. For another example, any and / or all functional modules in this application may be implemented in whole or in part by a shared processor executing software instructions. In addition, various software submodules executed by one or more processors may be shared between various software modules. Accordingly, unless explicitly required, the scope of this application is not limited by the mandatory boundaries between various hardware and / or software elements.

[0110] It should be noted that the order of description of the embodiments of the present application does not limit the priority order of the embodiments.

[0111] Unless otherwise defined, all technical and scientific terms used in this application have the same meaning as commonly understood by those skilled in the art to which this application belongs. The terms used in this application and in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application.

[0112] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of this application, ordinary technicians in this field can also make many forms without departing from the purpose of this application and the scope of protection of the claims. All equivalent transformations made by using the contents of the description and drawings of this application under the inventive concept of this application, or direct / indirect application in other related technical fields are included in the scope of patent protection of this application.

Claims

1. A method for detecting slitting defects of battery pole pieces, characterized in that: include: Performing a first process on the image to be identified of the battery electrode to obtain a first defect feature, where the area of ​​the first defect feature is greater than a first area threshold; performing a second process on the image to be identified to obtain a second defect feature, where an area of ​​the second defect feature is greater than a second area threshold and smaller than the first area threshold, and the second area threshold is smaller than the first area threshold; Based on the first defect feature and the second defect feature, an abnormal defect feature is determined, and the abnormal defect feature is marked.

2. The method according to claim 1, characterized in that The first processing is implemented based on an image processing method that does not rely on sample training.

3. The method according to claim 2, characterized in that The first process includes: After the image to be identified is binarized, a binary image is obtained; Scanning the area to be identified in the binary image and marking it according to the continuity of pixels to obtain at least two connected areas; The area of ​​each of the connected regions is calculated, and a connected region having an area greater than the first area threshold is identified as a first defect feature.

4. The method according to claim 1, wherein The second processing is implemented based on a neural network model trained based on defect samples.

5. The method according to claim 4, characterized in that The method further comprises: Based on the image data in which the second defect feature has been confirmed, the neural network model is incrementally trained to dynamically update the neural network model.

6. The method according to claim 1, characterized in that The method further comprises: Classifying the abnormal defect characteristics to obtain a third defect characteristic corresponding to the first defect type and a fourth defect characteristic corresponding to the second defect type; When the number of the third defect characteristics detected per unit length of the electrode exceeds a preset third threshold, a first adjustment signal is sent to the die-cutting parameter adjustment system to adjust the size of the slitting blade gap; When it is detected that the fourth defect feature appears continuously for a preset number of times in the pole piece running direction, a second adjustment signal is sent to the die-cutting parameter adjustment system to adjust the size of the feed amount.

7. The method according to claim 6, characterized in that The third defect feature is a feature corresponding to a powder falling defect, and the fourth defect feature is a feature corresponding to a burr defect.

8. The method according to claim 1, characterized in that The method further comprises: When a pole piece having the abnormal defect feature is detected, the marking device is controlled to set a defect mark at the location of the abnormal defect feature; Based on the defect mark, the pole pieces having the abnormal defect characteristics are eliminated.

9. The method according to claim 8, characterized in that The step of removing the electrode having the abnormal defect characteristics based on the defect mark includes: Acquire the electrode running length information in real time, and determine the position of the electrode to be rejected based on the electrode running length information and the identification information of the defect mark; When the interval between the consecutive electrode pieces to be removed is less than a preset length threshold, the consecutive electrode pieces to be removed and the electrode pieces in between are removed together; When the intervals between the consecutive electrode pieces to be removed are greater than the preset length threshold, only the electrode pieces to be removed are removed.

10. A slitting defect detection control device, characterized in that: include: A first processing module is configured to perform a first processing on the image to be identified of the battery electrode to obtain a first defect feature, wherein the area of ​​the first defect feature is greater than a first area threshold; a second processing module, configured to perform a second processing on the image to be identified to obtain a second defect feature, wherein an area of ​​the second defect feature is greater than a second area threshold and smaller than the first area threshold, and the second area threshold is smaller than the first area threshold; The abnormal defect marking module is used to determine an abnormal defect feature based on the first defect feature and the second defect feature, and mark the abnormal defect feature.

11. A slitting defect detection and control device, characterized in that: The method comprises a memory and a processor, wherein the memory is used to store computer programs or instructions; when the computer program or instructions are executed by the processor, the method according to any one of claims 1 to 9 is implemented.

12. A slitting defect detection and control system, characterized in that: include: The device according to claim 10 or claim 11; The image acquisition module is connected to the device and is used to acquire images of the battery pole pieces after cutting in real time, obtain images to be identified and send them to the device.

13. The system according to claim 12, wherein: The image acquisition module includes a liquid lens camera with dynamic focus or depth of field adjustment function.

14. A computer-readable storage medium, characterized in that The storage medium stores a computer program or instruction, and when the computer program or instruction is executed by a processor, the method according to any one of claims 1 to 9 is implemented.

15. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed by a processor, the method according to any one of claims 1 to 9 is implemented.

Images