Battery production defect inspection method, system, medium and equipment

Abstract

The invention discloses a battery production defect inspection method, system, medium and equipment, and relates to the technical field of battery defect inspection. The method comprises the following steps: acquiring a video stream corresponding to each battery production line, and extracting a key frame sequence corresponding to each video stream; performing fusion analysis on spatial features and time sequence features of the key frame sequences to obtain abnormal state features; performing multi-stage screening on each key frame sequence based on the abnormal state features to obtain a target key frame with defects; performing space-time mapping on the target key frame, and determining an associated station corresponding to the abnormal production line; and according to the process parameters of the associated stations, carrying out grading evaluation on the abnormal state features, and generating abnormal early warning information. By implementing the technical scheme provided by the invention, misjudgment or missed judgment of abnormal conditions on the battery production line can be effectively avoided, so that the production quality of battery products is improved.

Description

Technical Field

[0001] The present application relates to the technical field of battery defect inspection, and in particular to a battery production defect inspection method, system, medium and equipment. Background Art

[0002] The battery production process involves multiple steps, complex production technology, and strict requirements on product quality. Therefore, timely discovery and resolution of defects in the production process is of great significance to ensuring battery product quality and improving production efficiency.

[0003] At present, battery production lines usually use a combination of video monitoring and manual inspections for quality inspection. Operators observe video monitoring images and combine them with process parameter displays to identify and handle abnormal situations that may occur during the production process. However, since battery production lines are usually large in scale and have many processes, it is difficult to achieve real-time monitoring of all workstations by relying solely on manual observation, which can easily lead to misjudgment or omission of abnormal situations on paired battery production lines, thereby reducing the production quality of battery products. Summary of the invention

[0004] The present application provides a battery production defect inspection method, which can effectively avoid misjudgment or omission of abnormal conditions on a battery production line, thereby improving the production quality of battery products.

[0005] In a first aspect, the present application provides a method for inspecting defects in battery production, the method comprising: Obtaining the video stream corresponding to each battery production line, and extracting the key frame sequence corresponding to each of the video streams; Performing a fusion analysis of spatial features and temporal features on each of the key frame sequences to obtain abnormal state features; Performing multi-level screening on each of the key frame sequences based on the abnormal state characteristics to obtain target key frames with defects; Performing spatiotemporal mapping on the target key frame to determine the associated workstations corresponding to the abnormal production line; According to the process parameters of the associated workstations, the abnormal state characteristics are graded and evaluated, and abnormal warning information is generated.

[0006] By adopting the above technical solution, by acquiring the video stream of the battery production line and extracting the key frame sequence, it is possible to replace manual observation to achieve continuous monitoring of all workstations; the key frame sequence is subjected to a fusion analysis of spatial features and temporal features, combined with a multi-level screening mechanism, to overcome the misjudgment problem that is easy to occur in manual observation, and accurately identify the target key frames with defects; by performing spatiotemporal mapping of the target key frames, the associated workstations of the abnormal production line can be quickly located, and the abnormal state characteristics can be graded and evaluated in combination with the process parameters of the workstation, thereby generating abnormal warning information. This automated defect inspection method not only solves the problem that manual monitoring is difficult to cover all workstations, but also can effectively avoid misjudgments or missed judgments caused by manual observation through a systematic feature analysis and evaluation mechanism, thereby improving the production quality of battery products.

[0007] Optionally, obtaining the video streams corresponding to each battery production line and extracting the key frame sequences corresponding to each of the video streams includes: Determine the corresponding video acquisition parameters according to the workstation type of each battery production line; Collect the video stream corresponding to each of the battery production lines based on the video acquisition parameters; Preprocessing each of the video streams to obtain a standard video stream; A corresponding key frame sequence is extracted from each of the standard video streams.

[0008] Optionally, extracting a corresponding key frame sequence from each of the standard video streams includes: Obtaining the process operation specifications corresponding to each of the battery production lines, and extracting the operation steps and timing requirements of each process in the process operation specifications; According to the operation steps, dividing a process detection window in the standard video stream; Identifying the state change process of the operation object in each process detection window; Determine the critical operation moment in combination with the timing requirements and the state change process; The corresponding process status frames are extracted at the key operation moments, and the process status frames are organized into key frame sequences according to the process flow.

[0009] Optionally, the performing a fusion analysis of spatial features and temporal features on each of the key frame sequences to obtain abnormal state features includes: Constructing a process feature library, wherein the process feature library includes standard spatial features and temporal features of each process in each battery production line; Performing spatiotemporal analysis on the key frame sequence to extract workpiece position features and operation timing features; Calculating the deviation between the workpiece position feature and the standard space feature, and calculating the difference between the operation timing feature and the standard timing feature; Constructing a feature fusion matrix based on the deviation value and the difference; The abnormal state features are identified according to the feature fusion matrix.

[0010] Optionally, the performing multi-level screening on each of the key frame sequences based on the abnormal state feature to obtain a target key frame with defects includes: Establishing a hierarchical screening rule base according to the type of the abnormal state characteristics; Performing a preliminary screening on the key frame sequence based on the hierarchical screening rule base to obtain a first abnormal frame set; Performing adjacent frame correlation analysis on the first abnormal frame set to obtain a second abnormal frame set; Verify and screen the second-level abnormal frame set based on the process state transition rule; The keyframes that pass the validation filter are determined as target keyframes.

[0011] Optionally, performing spatiotemporal mapping on the target key frame to determine the associated workstation corresponding to the abnormal production line includes: Constructing a topological structure diagram of the workstations corresponding to each of the battery production lines; Extracting process timestamps and spatial coordinates from the target keyframe; Based on the process timestamp and the spatial coordinates, the associated workstations corresponding to the abnormal production line are located in the workstation topology structure diagram.

[0012] Optionally, the step of performing a graded evaluation on the abnormal state characteristics according to the process parameters of the associated workstations and generating abnormal warning information includes: Obtaining the process parameters of the associated workstation and the corresponding process standard range; Calculating the deviation of the process parameters from the process standard range; Quantitatively scoring the abnormal state feature based on the deviation, and determining the abnormality level according to the quantitative score; The impact of the abnormality level on the battery production line is analyzed, and graded abnormality warning information including the abnormality level and the impact level is generated.

[0013] In a second aspect of the present application, a battery production defect inspection system is provided, the system comprising: A frame sequence extraction module, used to obtain the video stream corresponding to each battery production line, and extract the key frame sequence corresponding to each video stream; A feature analysis module is used to perform a fusion analysis of spatial features and temporal features on each of the key frame sequences to obtain abnormal state features; A defect detection module, used for performing multi-level screening on each of the key frame sequences based on the abnormal state characteristics to obtain target key frames with defects; An abnormality determination module, used to perform spatiotemporal mapping on the target key frame and determine the associated workstations corresponding to the abnormal production line; The abnormal warning module is used to perform graded evaluation on the abnormal state characteristics according to the process parameters of the associated workstations and generate abnormal warning information.

[0014] In a third aspect of the present application, a computer storage medium is provided, wherein the computer storage medium stores a plurality of instructions, wherein the instructions are suitable for being loaded by a processor and executing the above method steps.

[0015] In a fourth aspect of the present application, an electronic device is provided, comprising: a processor and a memory; wherein the memory stores a computer program, and the computer program is suitable for being loaded by the processor and executing the above-mentioned method steps.

[0016] In summary, one or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: The present application can replace manual observation to achieve continuous monitoring of all workstations by acquiring the video stream of the battery production line and extracting the key frame sequence; the key frame sequence is subjected to a fusion analysis of spatial features and temporal features, combined with a multi-level screening mechanism, to overcome the misjudgment problem that is easy to be caused by manual observation, and accurately identify target key frames with defects; by performing spatiotemporal mapping of the target key frames, the associated workstations of the abnormal production line can be quickly located, and the abnormal state characteristics can be graded and evaluated in combination with the process parameters of the workstation, thereby generating abnormal warning information. This automated defect inspection method not only solves the problem that manual monitoring is difficult to cover all workstations, but also can effectively avoid misjudgments or missed judgments caused by manual observation through a systematic feature analysis and evaluation mechanism, thereby improving the production quality of battery products. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a flow chart of a battery production defect inspection method provided in an embodiment of the present application; Figure 2 It is a module schematic diagram of a battery production defect inspection system provided in an embodiment of the present application; Figure 3 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application.

[0018] Description of reference numerals: 300, electronic device; 301, processor; 302, communication bus; 303, user interface; 304, network interface; 305, memory. DETAILED DESCRIPTION

[0019] In order to enable technicians in this field to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments.

[0020] In the description of the embodiments of the present application, words such as "for example" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "for example" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "for example" or "for example" is intended to present related concepts in a specific way.

[0021] In the description of the embodiments of the present application, the meaning of the term "multiple" refers to two or more. For example, multiple systems refer to two or more systems, and multiple screen terminals refer to two or more screen terminals. In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. The terms "include", "comprise", "have" and their variations all mean "including but not limited to", unless otherwise specifically emphasized.

[0022] The following will provide a clear and complete description of the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments.

[0023] Please refer to Figure 1 , a flowchart of a battery production defect inspection method is proposed, the method can be implemented by a computer program, can be implemented by a single-chip microcomputer, and can also be run on a battery production defect inspection system. The computer program can be integrated in a computer device or run as an independent tool application. Specifically, the method includes steps 10 to 50, and the above steps are as follows: Step 10: Obtain the video stream corresponding to each battery production line, and extract the key frame sequence corresponding to each video stream.

[0024] The battery production line is an automated production line for producing batteries in the embodiment of the present application, including multiple process stations such as electrode preparation, winding, liquid injection, and formation. The stations are connected by a transmission device to form a complete production line.

[0025] In the embodiment of the present application, the video stream refers to the continuous video data stream of each workstation of the battery production line collected by the video monitoring equipment, and the video data stream records the production process and equipment operation status of each workstation in real time.

[0026] In the embodiment of the present application, the key frame sequence refers to a set of image frames representing the key states of the process extracted from the video stream according to the process operation specifications. These key frames reflect the important operation moments and workpiece state changes in each process, and are arranged and organized in the order of the process flow.

[0027] Specifically, since different workstations on the battery production line have different process characteristics and monitoring requirements, it is first necessary to determine the corresponding video acquisition parameters, such as sampling frequency, resolution, etc., according to the workstation type of each battery production line. Based on the determined video acquisition parameters, the video stream of the battery production line is collected by the video monitoring equipment deployed at each workstation. In order to ensure the accuracy of subsequent analysis, the collected video stream needs to be preprocessed, including noise elimination, brightness and contrast adjustment, etc., to obtain a standardized video stream. Obtain the process operation specifications corresponding to each battery production line, and extract the standard operation steps and timing requirements of each process from them. According to the standard operation steps, divide the process detection window in the standard video stream, and identify the state change process of the work object in each process detection window. Based on the identified state change process, the key operation moments are determined in combination with the timing requirements of the process, and the corresponding process state frames are extracted at these key moments. Finally, the extracted process state frames are organized into a key frame sequence according to the process flow sequence.

[0028] Based on the above embodiment, as an optional embodiment, the step of obtaining the video stream corresponding to each battery production line and extracting the key frame sequence corresponding to each video stream may also include the following steps: Step 101: Determine corresponding video acquisition parameters according to the workstation type of each battery production line.

[0029] Step 102: Capture the video stream corresponding to each battery production line based on the video acquisition parameters.

[0030] Specifically, due to the differences in production characteristics and monitoring requirements of different workstations on the battery production line, for example, the pole piece preparation workstation needs to pay attention to coating uniformity, and the winding workstation needs to pay attention to winding tightness, so it is necessary to set the video acquisition parameters in a targeted manner. Specifically, first establish a mapping relationship table between workstation type and video acquisition parameters, which contains parameter configurations such as image resolution, sampling frame rate, and field of view angle required for different workstations. For workstations that need to observe fast-moving details, set a higher sampling frame rate; for workstations that need to identify subtle features, set a higher image resolution; for workstations with larger spaces, set a suitable field of view angle to ensure monitoring coverage. According to the mapping relationship table, configure the corresponding acquisition parameters for the video monitoring equipment of each workstation, and collect the video stream of each battery production line in real time based on the configured acquisition parameters. This differentiated parameter configuration method not only ensures that the collected video data can meet the monitoring needs of each workstation, but also avoids data redundancy or information loss caused by improper parameter configuration, thereby providing reliable quality raw data for subsequent anomaly detection.

[0031] Step 103: Preprocess each video stream to obtain a standard video stream.

[0032] Specifically, since the collected original video stream may be affected by factors such as uneven lighting and device jitter, preprocessing is required to improve the accuracy of subsequent analysis. First, the video stream is denoised, and algorithms such as Gaussian filtering are used to remove random noise; then lighting compensation and contrast enhancement are performed to make the image details clearer; finally, image stabilization is performed to eliminate the image jitter caused by device vibration, thereby obtaining a standardized video stream.

[0033] Step 104: extracting corresponding key frame sequences from each standard video stream.

[0034] Specifically, in order to accurately extract the key frame sequence that reflects the process status, it is first necessary to obtain the process operation specifications corresponding to each battery production line. The standard operating steps and timing requirements of each process (such as electrode preparation, winding, injection, etc.) are extracted from the process operation specifications. The operating steps describe the specific execution process of the process, and the timing requirements specify the execution time and sequence of each step. Based on the extracted operating step information, the corresponding process detection windows are divided on the timeline of the standard video stream. Each detection window corresponds to a complete process, such as the entire process from the start of electrode coating to the completion of coating.

[0035] In the divided process detection window, the state change process of the working object (such as pole piece, diaphragm, electrolyte, etc.) is identified through target detection and tracking algorithms. Specifically, the position coordinates, morphological characteristics, motion trajectory and other parameters of the working object are tracked in real time, and the change trend of these parameters over time is recorded. Combined with the timing requirements in the process operation specifications, the key operation moments are determined based on the state change process of the working object, such as the pole piece entering the coating area, the coating slurry contacting the pole piece, the coating completion and other important time points.

[0036] At the determined critical operation moment, the corresponding process status frames are extracted from the standard video stream. These status frames clearly record the status information of the work object at the critical moment, such as coating uniformity, winding tightness, etc. Finally, the extracted process status frames are organized into a key frame sequence according to the sequence of the process flow to form a complete process status record. This key frame extraction method based on process specification guidance ensures that the extracted frame sequence can accurately reflect the critical status during the process execution, providing reliable analysis data for subsequent anomaly detection. At the same time, by selectively extracting key frames, the amount of data processing is greatly reduced.

[0037] Step 20: Perform spatial feature and temporal feature fusion analysis on each key frame sequence to obtain abnormal state features.

[0038] Specifically, since battery production defects may be manifested in two dimensions, namely, spatial features (such as uneven coating, material deformation, etc.) and temporal features (such as abnormal process execution timing, abnormal state transition, etc.), it is necessary to fuse and analyze these two types of features to comprehensively identify abnormal states. First, spatial features are extracted from each frame image in the key frame sequence, including morphological features of the work object (such as the shape and size of the pole piece), surface features (such as the uniformity and defects of the coating), and position features (such as relative position, posture angle, etc.). Then, based on the extracted spatial features, a deep learning model is used to evaluate the state of the work object and obtain the spatial domain abnormality index. At the same time, the key frame sequence is analyzed in a temporal order to extract the temporal features during the process execution, including dynamic parameters such as the time interval of the state transition of the work object, the speed and acceleration of the state change. The changing rules of these dynamic parameters are analyzed through the temporal model, the temporal rationality of the process execution is evaluated, and the temporal domain abnormality index is obtained. In order to comprehensively consider the abnormal features of the two dimensions of space and time, the feature fusion algorithm is used to weightedly fuse the spatial domain abnormality index and the temporal domain abnormality index to construct a multi-dimensional abnormal state feature.

[0039] Based on the above embodiment, as an optional embodiment, the step of performing a fusion analysis of spatial features and temporal features on each key frame sequence to obtain abnormal state features may also include the following steps: Step 201: construct a process feature library, which includes standard spatial features and temporal features of each process in each battery production line.

[0040] Specifically, in order to accurately judge the abnormal state in the battery production process, it is necessary to first establish a standard process feature library as a reference benchmark. By collecting historical data of normal operation of each battery production line, the standard features of each process are extracted, including spatial features (such as standard coating thickness distribution during electrode coating, standard winding diameter during winding, etc.) and timing features (such as standard execution time of coating process, standard injection rate of injection process, etc.), and these feature data are classified and stored according to process type to form a complete process feature library.

[0041] Step 202: Perform spatiotemporal analysis on the key frame sequence to extract workpiece position features and operation timing features.

[0042] Specifically, on the basis of constructing the process feature library, the key frame sequence is analyzed in time and space to obtain the actual features of the current production process. In the spatial dimension, the position features of the workpiece are extracted through image segmentation and target detection algorithms, including the spatial coordinates, posture angles, morphological parameters, etc. of the workpiece on the production line; in the temporal dimension, the operation timing features are extracted by analyzing the state changes between adjacent key frames, including the transmission time, processing time, and state conversion time of the workpiece between each process. This time and space analysis method can not only accurately locate the spatial position and state of the workpiece on the production line, but also reflect the temporal laws of the process execution process, providing complete feature data for subsequent comparative analysis with the standard features in the process feature library, thereby effectively improving the accuracy of anomaly detection.

[0043] Step 203: Calculate the deviation between the workpiece position feature and the standard space feature, and calculate the difference between the operation timing feature and the standard timing feature.

[0044] Specifically, in order to quantitatively evaluate the abnormality of the workpiece status and process execution, it is necessary to compare and analyze the actual features extracted with the standard features in the process feature library. First, calculate the deviation value between the workpiece position feature and the standard spatial feature. Specifically, for the spatial coordinates, posture angle, morphological parameters and other features of the workpiece, calculate the Euclidean distance from the standard value, and set the weight coefficient according to the importance of each feature. The comprehensive deviation value is obtained through weighted calculation. For example, for the electrode coating process, calculate the deviation between the coating thickness distribution and the standard distribution, as well as the degree of deviation between the electrode position and the standard processing position.

[0045] At the same time, the difference between the operation timing characteristics and the standard timing characteristics is calculated. By comparing the timing parameters such as the transmission time, processing time, and state conversion time of the workpiece between each process with the standard value, the difference indicators such as time delay rate and beat deviation are calculated. In order to comprehensively reflect the degree of timing anomaly, the dynamic time warping algorithm is used to calculate the matching degree between the actual timing characteristics and the standard timing characteristics to obtain the timing difference. This quantitative analysis method based on spatiotemporal characteristics can not only accurately measure the degree of abnormality of the workpiece state and process execution, but also highlight the influence of key features through weight configuration, providing a reliable quantitative basis for the subsequent abnormal state feature fusion, and effectively improving the accuracy of anomaly detection.

[0046] Step 204: construct a feature fusion matrix based on the deviation value and the difference.

[0047] Specifically, in order to comprehensively consider the abnormal features of spatial and temporal dimensions, it is necessary to construct a feature fusion matrix to achieve effective fusion of multi-dimensional features. The calculated workpiece position feature deviation values ​​and operation timing feature differences are arranged in the order of the processes to form a two-dimensional feature matrix. The rows of the matrix represent different processes, and the columns represent spatial features and temporal features. The elements in the feature fusion matrix are normalized to make features of different dimensions comparable, and the corresponding weight coefficients are set according to the importance of each process and the degree of influence on product quality.

[0048] Step 205: Identify abnormal state features according to the feature fusion matrix.

[0049] Specifically, based on the feature fusion matrix, a deep learning model is used to analyze the matrix data and identify abnormal state features. The local and global features in the matrix are extracted through a convolutional neural network to capture the correlation between processes and the law of abnormal transmission. The abnormal state features output by the model include information such as abnormal type, abnormal degree and abnormal location. The abnormal recognition method based on the feature fusion matrix effectively organizes multi-dimensional abnormal features in the form of a matrix, and uses a deep learning model to mine the potential correlation between features, which not only improves the accuracy of abnormal detection, but also reflects the transmission effect of abnormal states between processes, providing a more comprehensive decision-making basis for subsequent abnormal warning and processing.

[0050] Step 30: Perform multi-level screening on each key frame sequence based on the abnormal state characteristics to obtain target key frames with defects.

[0051] Specifically, in order to accurately locate the key frames with defects and reduce the misjudgment rate, it is necessary to perform multi-level screening on the key frame sequence. First, a preliminary screening is performed based on the degree of abnormality in the abnormal state characteristics, and the key frames with abnormality levels exceeding the preset threshold are screened out to form a preliminary frame set. For the preliminary frame set, a secondary screening is further performed in combination with the abnormal type characteristics. The defect characteristics in each key frame are finely classified through a deep learning model to identify whether the defect type belongs to the predefined category of concern defects, and frames that do not meet the requirements are eliminated. For the key frames that pass the secondary screening, a tertiary screening is performed based on the abnormal position characteristics to analyze the distribution position and transmission relationship of the defects in the process flow, and identify whether the defects will affect subsequent processes. For example, abnormal electrode coating uniformity may affect the product quality of the subsequent winding process. This type of defective frame with a transmission effect needs to be paid special attention to. Finally, the key frames that pass the multi-level screening are determined as target key frames with defects. This multi-level screening method, through gradually refined screening criteria, not only ensures the accuracy of defect detection, but also improves the screening efficiency. Based on the above embodiment, as an optional embodiment, the step of performing multi-level screening on each key frame sequence based on abnormal state characteristics to obtain a target key frame with defects may also include the following steps: Step 301: Establish a hierarchical screening rule base according to the type of abnormal state characteristics.

[0052] Specifically, in order to achieve accurate identification and hierarchical control of abnormal conditions, it is first necessary to establish a hierarchical screening rule base based on the type of abnormal state characteristics. Based on historical production data and expert experience, abnormal state characteristics can be divided into different levels according to the degree of impact, such as severe abnormalities, minor abnormalities, etc., and corresponding screening rules can be formulated for different levels. For example, for coating thickness abnormalities in the coating process, different alarm thresholds can be set according to the degree of deviation; for tension abnormalities in the winding process, different screening criteria can be determined according to the fluctuation amplitude. These rules are systematically stored in the hierarchical screening rule base to form a complete screening system.

[0053] Step 302: Preliminarily screen the key frame sequence based on the hierarchical screening rule base to obtain a first abnormal frame set.

[0054] Specifically, the abnormal state features are extracted for each key frame, including the workpiece position feature deviation value and the operation timing feature difference, and these feature values ​​are compared with the thresholds in the hierarchical screening rule library. For example, when the coating thickness deviation in the electrode coating process exceeds the threshold range set by the rule library, or the process execution time delay exceeds the allowable fluctuation range of the standard timing, the corresponding key frame is marked as an abnormal frame. At the same time, taking into account the combined effect of abnormal states, when multiple low-level abnormal features appear at the same time and their combination pattern conforms to the composite abnormal rule in the rule library, the corresponding frames are also included in the abnormal frame set. Through this multi-dimensional feature comparison and rule matching, the first abnormal frame set is obtained.

[0055] Step 303: performing adjacent frame correlation analysis on the first abnormal frame set to obtain a second abnormal frame set.

[0056] Specifically, in order to deeply analyze the propagation law and impact range of abnormal states, it is necessary to perform adjacent frame correlation analysis on the frames in the first abnormal frame set. First, calculate the state feature correlation between adjacent frames in the first abnormal frame set, including the continuous change of spatial features and the gradual change of temporal features. When it is detected that the previous and next frames of an abnormal frame have not reached the initial screening abnormal threshold, but have significant feature correlation with the abnormal frame, these related frames are also included in the analysis scope. For example, around the key frame where the coating uniformity anomaly is found, the system will analyze the change trend of the coating parameters in the adjacent frames. If the cumulative effect or attenuation process of abnormal signs is found, these related frames will be included in the second abnormal frame set. This analysis method based on the correlation of adjacent frames can not only track the evolution of abnormal states, but also promptly discover potential abnormal trends, thereby improving the foresight and completeness of anomaly detection.

[0057] Step 304: verify and screen the second-level abnormal frame set based on the process state transition rule.

[0058] Specifically, the second abnormal frame set is verified and screened based on the process state transition rules. The main purpose is to verify the rationality of the abnormal state through process logic and eliminate misjudgment. First, the process state transition rules are extracted from the process operation specifications. These rules define the reasonable transition paths and constraints between each process state. For example, in the electrode preparation process, the coating state must be started after the electrode cleaning is completed, and the change of coating parameters needs to comply with the preset gradual change law. The system converts these process logics into state transition rules to verify whether the abnormal state in the second abnormal frame set violates the process specifications.

[0059] Step 305: Determine the key frame that passes the verification screening as the target key frame.

[0060] Specifically, for each key frame in the second abnormal frame set, the state sequence of the process stage in which it is located is analyzed to verify the rationality of the state transition. First, check whether the abnormal state is in the allowed process stage, for example, some parameter adjustments can only be performed in a specific process stage. Then analyze whether the abnormal state conversion process meets the process specifications, including the timing, continuity and causality of the state conversion. For those state frames that show abnormal characteristics but actually meet the process adjustment specifications, they are removed from the abnormal frame set. Finally, the key frames that pass the verification screening are determined as target key frames. These frames not only show abnormalities at the feature level, but also have violations in process logic. The verification and screening method based on the process state transition rules effectively improves the accuracy of anomaly detection by introducing process logic constraints, avoids misjudgments caused by normal operations such as process adjustments, and ensures the consistency of the detection results with the actual production conditions.

[0061] Step 40: Perform spatiotemporal mapping on the target key frame to determine the associated workstations corresponding to the abnormal production line.

[0062] Specifically, in order to accurately locate the specific location where the abnormal state occurs and analyze its impact range, it is necessary to perform spatiotemporal mapping on the target key frame. First, based on the timestamp information in the target key frame, the abnormal state is mapped to the actual operating sequence of the production line to determine the time node when the abnormality occurs. At the same time, the spatial feature information in the target key frame, including the workpiece position feature and the equipment status feature, is used in combination with the layout information of the production line to map the abnormal state to a specific station coordinate system. On the basis of spatiotemporal mapping, the association relationship between the abnormal station and the adjacent station is further analyzed. By analyzing the process flow chart and the equipment layout diagram, the associated stations with material transfer, process connection or equipment linkage relationships with the abnormal station are identified. For example, when an abnormality is detected in the electrode coating station, the system will automatically analyze the status of the upstream feeding station and the downstream receiving station of the station to determine whether the abnormality will affect the normal operation of these associated stations. This association analysis method based on spatiotemporal mapping can not only accurately locate the specific location where the abnormality occurs, but also predict the propagation path of the abnormal state, providing spatial guidance for timely and targeted intervention measures, and effectively improving the accuracy and efficiency of abnormality handling.

[0063] Based on the above embodiment, as an optional embodiment, the step of performing spatiotemporal mapping on the target key frame to determine the associated workstation corresponding to the abnormal production line may further include the following steps: Step 401: construct a workstation topology diagram corresponding to each battery production line.

[0064] Specifically, the equipment layout data and process flow data of each battery production line are collected, including the spatial coordinates of the workstation, equipment size, material transmission path and other information. Based on these basic data, a graph structure representation method is adopted, each workstation is used as a node in the graph, and the material transmission path and process association between the workstations are used as connecting edges to construct an initial topological network. During the construction process, the process attributes of the nodes are further annotated, such as the workstation type, processing capacity, equipment status, etc., and weight attributes are assigned to the connecting edges, including material transmission distance, process connection time, etc. At the same time, the dependencies between nodes are annotated according to the process requirements. For example, there is a strict process sequence between the coating station and the rolling station. For parallel stations, their logical associations are represented by adding virtual connecting edges. This visual representation based on the topological structure of the station not only intuitively shows the spatial layout and process flow of the production line, but also reflects the complex association between the stations through the attribute annotation of the graph structure, providing a systematic reference model for subsequent abnormal propagation analysis and process optimization.

[0065] Step 402: extracting the process timestamp and spatial coordinates from the target key frame.

[0066] Specifically, in order to accurately locate the propagation range of abnormal conditions in the production line, it is necessary to extract key information from the target key frame, and extract the timestamp information of the process execution by parsing the metadata of the target key frame, including the start time, duration and end time of the abnormal condition; at the same time, extract the spatial coordinate information of the workpiece during processing, including the position, posture and motion trajectory parameters of the workpiece in the production line coordinate system. These spatiotemporal information serve as the basic data for locating abnormal workstations.

[0067] Step 403: Based on the process timestamp and the spatial coordinates, locate the associated workstations corresponding to the abnormal production line in the workstation topology diagram.

[0068] Specifically, based on the extracted process timestamps and spatial coordinates, the abnormality is located in the constructed workstation topology diagram. First, the abnormal state is mapped to the specific node in the topology diagram using spatial coordinates to determine the initial workstation where the abnormality occurs. Then, the timing characteristics of the abnormal state are analyzed based on the process timestamp, and the associated workstations that may be affected by the abnormal state are tracked in combination with the node connection relationship in the topology diagram. For example, when the workpiece position deviation of a certain workstation is detected, the degree of influence of the deviation on the downstream workstation can be analyzed along the material transmission path in the topology diagram; at the same time, the abnormal state can be judged according to the process timestamp whether it will affect the processing rhythm of the parallel workstations. This method of locating associated workstations based on spatiotemporal characteristics can not only accurately identify the spatial distribution of abnormalities, but also predict the propagation trend of abnormalities by establishing a corresponding relationship between abnormal states and workstations in the topology diagram.

[0069] Step 50: Based on the process parameters of the associated workstations, the abnormal state characteristics are graded and evaluated, and abnormal warning information is generated.

[0070] Specifically, the key process parameters of the associated stations are extracted, including process setting values, allowable fluctuation range, quality control standards and other information, and these parameters are matched and analyzed with the abnormal state characteristics. For example, for the thickness abnormality of the coating station, the system will combine the pressure parameters of the downstream rolling station and the thickness requirements of the pole piece to evaluate the degree of influence of the abnormal state on the product quality. On the basis of parameter matching, the system uses a multi-level evaluation model to classify the abnormal state. First, the basic level is determined according to the degree of deviation between the abnormal characteristics and the process standards, and then the level is adjusted considering the propagation effect of the abnormal state between the associated stations. For example, when an abnormal state may cause the process parameters of multiple associated stations to exceed the standard, the system will increase the abnormal level. For the evaluation results, the system automatically generates early warning information containing abnormal level, impact range, processing suggestions, etc., and selects different early warning methods according to the abnormal level, such as real-time pop-up windows, sound and light alarms, etc. This hierarchical evaluation method based on process parameters, by establishing a quantitative relationship between abnormal state and quality impact, not only improves the accuracy and pertinence of early warning, but also provides clear priority guidance for abnormal handling, effectively improving the rapid response capability and quality control level of the production line.

[0071] Based on the above embodiment, as an optional embodiment, the step of performing graded evaluation on abnormal state characteristics according to the process parameters of the associated workstations and generating abnormal warning information may further include the following steps: Step 501: Obtain the process parameters of the associated workstations and the corresponding process standard range.

[0072] Specifically, the real-time process parameters of the related stations are extracted from the process database, including the coating thickness of the coating station, the pressure value of the rolling station, the size parameters of the cutting station and other key process data. At the same time, the process standard ranges corresponding to these stations are retrieved, including the target values ​​of the parameters, the upper and lower limit thresholds and the allowed fluctuation range. These standard ranges are determined based on product quality requirements and process experience.

[0073] Step 502: Calculate the deviation between the process parameters and the process standard range.

[0074] Specifically, a multi-dimensional deviation calculation method is used to analyze the acquired process parameters. First, the absolute deviation between the process parameters and the standard target values ​​is calculated to evaluate the discreteness of the parameters; then the fluctuation trend of the parameters is calculated to analyze whether there are gradual or sudden changes; finally, combined with the process standard range, the degree and duration of the parameters exceeding the allowable range are calculated. For example, when a coating thickness deviation is detected, not only the difference between the thickness value and the standard value is calculated, but also the uniformity and change trend of the thickness distribution are analyzed. This multi-dimensional deviation calculation method accurately reflects the abnormality of the process parameters through quantitative data analysis, provides an objective basis for subsequent graded evaluation, and effectively improves the accuracy of abnormality identification.

[0075] Step 503: quantitatively score the abnormal state feature based on the deviation, and determine the abnormal level according to the quantitative score.

[0076] Specifically, based on the calculated process parameter deviation value, a weighted scoring method is used to quantitatively score the abnormal state characteristics. For example, for coating thickness abnormality, different weight coefficients are set to calculate the comprehensive score of the abnormal state by comprehensively considering factors such as the deviation value, duration, and change trend. When the deviation of a parameter exceeds the standard range, the scoring weight is increased according to the degree of excess and duration; when multiple parameters are abnormal at the same time, the coupling effect between the parameters is considered for scoring superposition. Based on the quantitative scoring results, the system divides the abnormal state into different levels according to the preset scoring interval. For example, the score is divided into three levels: slight abnormality, general abnormality, and severe abnormality, and each level corresponds to a different score range and processing requirements. Furthermore, the system analyzes the degree of influence of each level of abnormality on the battery production line, including the direct impact on product quality, the chain effect on subsequent processes, and the overall impact on production efficiency. For example, when a serious abnormality is detected in a coating station, the system will evaluate the quality problems such as decreased battery capacity and reduced yield rate that may be caused by the abnormality, as well as the loss of production capacity due to process stagnation.

[0077] Step 504: Analyze the impact of the abnormality level on the battery production line, and generate graded abnormality warning information including the abnormality level and the impact level.

[0078] Specifically, based on the analysis results, the system automatically generates graded abnormal warning information, which includes key content such as abnormal workstation, abnormal level, impact range, and handling suggestions. Different warning methods and handling processes are used for different levels of abnormalities. For example, for minor abnormalities, the system generates reminder information and recommends that relevant personnel pay attention; for serious abnormalities, an alarm signal is immediately triggered and the relevant responsible person is automatically notified for processing. This graded warning mechanism based on quantitative scoring not only improves the accuracy and timeliness of the warning by establishing a quantitative relationship between abnormal status and production impact, but also provides clear priority guidance for abnormal handling, effectively improving the abnormal control capability and quality assurance level of the production line.

[0079] See also Figure 3 , is a module diagram of a battery production defect inspection system provided in an embodiment of the present application, the battery production defect inspection system may include: a frame sequence extraction module, a feature analysis module, a defect detection module, an abnormality determination module and an abnormality warning module, wherein: A frame sequence extraction module, used to obtain the video stream corresponding to each battery production line, and extract the key frame sequence corresponding to each video stream; A feature analysis module is used to perform a fusion analysis of spatial features and temporal features on each of the key frame sequences to obtain abnormal state features; A defect detection module, used for performing multi-level screening on each of the key frame sequences based on the abnormal state characteristics to obtain target key frames with defects; An abnormality determination module, used to perform spatiotemporal mapping on the target key frame and determine the associated workstations corresponding to the abnormal production line; The abnormality warning module is used to perform graded evaluation on the abnormal state characteristics according to the process parameters of the associated workstations and generate abnormality warning information.

[0080] Optionally, the frame sequence extraction module is further used to determine corresponding video acquisition parameters according to the workstation type of each battery production line; Collect the video stream corresponding to each of the battery production lines based on the video acquisition parameters; Preprocessing each of the video streams to obtain a standard video stream; A corresponding key frame sequence is extracted from each of the standard video streams.

[0081] Optionally, the frame sequence extraction module is further used to obtain the process operation specifications corresponding to each of the battery production lines, and extract the operation steps and timing requirements of each process from the process operation specifications; According to the operation steps, dividing a process detection window in the standard video stream; Identifying the state change process of the operation object in each process detection window; Determine the critical operation moment in combination with the timing requirements and the state change process; The corresponding process status frames are extracted at the key operation moments, and the process status frames are organized into key frame sequences according to the process flow.

[0082] Optionally, the feature analysis module is further used to construct a process feature library, wherein the process feature library includes standard spatial features and temporal features of each process in each battery production line; Performing spatiotemporal analysis on the key frame sequence to extract workpiece position features and operation timing features; Calculating the deviation between the workpiece position feature and the standard space feature, and calculating the difference between the operation timing feature and the standard timing feature; Constructing a feature fusion matrix based on the deviation value and the difference; The abnormal state features are identified according to the feature fusion matrix.

[0083] Optionally, the defect detection module is further used to establish a hierarchical screening rule base according to the type of the abnormal state feature; Performing a preliminary screening on the key frame sequence based on the hierarchical screening rule base to obtain a first abnormal frame set; Performing adjacent frame correlation analysis on the first abnormal frame set to obtain a second abnormal frame set; Verify and screen the second-level abnormal frame set based on the process state transition rule; The keyframes that pass the validation filter are determined as target keyframes.

[0084] Optionally, the abnormality determination module is further used to construct a workstation topology diagram corresponding to each of the battery production lines; Extracting process timestamps and spatial coordinates from the target keyframe; Based on the process timestamp and the spatial coordinates, the associated workstations corresponding to the abnormal production line are located in the workstation topology structure diagram.

[0085] Optionally, the abnormal warning module is further used to obtain the process parameters of the associated workstation and the corresponding process standard range; Calculating the deviation of the process parameters from the process standard range; Quantitatively scoring the abnormal state feature based on the deviation, and determining the abnormality level according to the quantitative score; The impact of the abnormality level on the battery production line is analyzed, and graded abnormality warning information including the abnormality level and the impact level is generated.

[0086] It should be noted that: when the system provided in the above embodiment realizes its functions, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the system and method embodiments provided in the above embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0087] An embodiment of the present application also provides a computer storage medium, which can store multiple instructions. The instructions are suitable for being loaded by a processor and executing a battery production defect inspection method of the above embodiment. The specific execution process can be found in the specific description of the above embodiment, which will not be repeated here.

[0088] Please refer to Figure 3 The application also discloses an electronic device. Figure 3 The electronic device 300 may include: at least one processor 301 , at least one network interface 304 , a user interface 303 , a memory 305 , and at least one communication bus 302 .

[0089] The communication bus 302 is used to realize the connection and communication between these components.

[0090] The user interface 303 may include a display screen (Display) and a camera (Camera). Optionally, the user interface 303 may also include a standard wired interface and a wireless interface.

[0091] The network interface 304 may optionally include a standard wired interface or a wireless interface (such as a WI-FI interface).

[0092] Among them, the processor 301 may include one or more processing cores. The processor 301 uses various interfaces and lines to connect various parts in the entire server, and executes various functions of the server and processes data by running or executing instructions, programs, code sets or instruction sets stored in the memory 305, and calling data stored in the memory 305. Optionally, the processor 301 can be implemented in at least one hardware form of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), and programmable logic array (Programmable Logic Array, PLA). The processor 301 can integrate one or a combination of a central processing unit (Central Processing Unit, CPU), a graphics processing unit (Graphics Processing Unit, GPU) and a modem. Among them, the CPU mainly processes the operating system, user interface and application programs; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 301, and it can be implemented separately through a chip.

[0093] Among them, the memory 305 may include a random access memory (Random Access Memory, RAM) and may also include a read-only memory (Read-Only Memory). Optionally, the memory 305 includes a non-transitory computer-readable storage medium. The memory 305 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 305 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data involved in the above-mentioned various method embodiments, etc. The memory 305 may optionally also be at least one storage device located away from the aforementioned processor 301. Refer to Figure 3 , the memory 305 as a computer storage medium may include an operating system, a network communication module, a user interface module and an application program of a battery production defect inspection method.

[0094] exist Figure 3In the electronic device 300 shown, the user interface 303 is mainly used to provide an input interface for the user and obtain the data input by the user; and the processor 301 can be used to call the application program storing a battery production defect inspection method in the memory 305, and when executed by one or more processors 301, the electronic device 300 executes one or more of the methods described in the above embodiments. It should be noted that for the aforementioned method embodiments, for the sake of simple description, they are all expressed as a series of action combinations, but those skilled in the art should know that the present application is not limited to the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required for the present application.

[0095] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0096] In the several embodiments provided in this application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0097] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0098] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0099] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a memory and includes several instructions for a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned memory includes: various media that can store program codes, such as USB flash drives, mobile hard drives, magnetic disks or optical disks.

[0100] The above is only an exemplary embodiment of the present disclosure and cannot be used to limit the scope of the present disclosure. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the disclosure of the specification and the truth of practice, those skilled in the art will easily think of other embodiments of the present disclosure.

[0101] This application is intended to cover any variation, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the art not described in the present disclosure. The description and examples are to be regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. A battery production defect inspection method, characterized in that: The method comprises: Obtaining the video stream corresponding to each battery production line, and extracting the key frame sequence corresponding to each of the video streams; Performing a fusion analysis of spatial features and temporal features on each of the key frame sequences to obtain abnormal state features; Performing multi-level screening on each of the key frame sequences based on the abnormal state characteristics to obtain target key frames with defects; Performing spatiotemporal mapping on the target key frame to determine the associated workstations corresponding to the abnormal production line; According to the process parameters of the associated workstations, the abnormal state characteristics are graded and evaluated, and abnormal warning information is generated.

2. The battery production defect inspection method according to claim 1, characterized in that: The obtaining of the video streams corresponding to the battery production lines and the extraction of the key frame sequences corresponding to the video streams include: Determine the corresponding video acquisition parameters according to the workstation type of each battery production line; Collect the video stream corresponding to each of the battery production lines based on the video acquisition parameters; Preprocessing each of the video streams to obtain a standard video stream; A corresponding key frame sequence is extracted from each of the standard video streams.

3. The battery production defect inspection method according to claim 2, characterized in that: The step of extracting a corresponding key frame sequence from each of the standard video streams comprises: Obtaining the process operation specifications corresponding to each of the battery production lines, and extracting the operation steps and timing requirements of each process in the process operation specifications; According to the operation steps, dividing a process detection window in the standard video stream; Identifying the state change process of the operation object in each process detection window; Determine the critical operation moment in combination with the timing requirements and the state change process; The corresponding process status frames are extracted at the key operation moments, and the process status frames are organized into key frame sequences according to the process flow.

4. The battery production defect inspection method according to claim 1, characterized in that: The performing of spatial feature and temporal feature fusion analysis on each of the key frame sequences to obtain abnormal state features includes: Constructing a process feature library, wherein the process feature library includes standard spatial features and temporal features of each process in each battery production line; Performing spatiotemporal analysis on the key frame sequence to extract workpiece position features and operation timing features; Calculating the deviation between the workpiece position feature and the standard space feature, and calculating the difference between the operation timing feature and the standard timing feature; Constructing a feature fusion matrix based on the deviation value and the difference; The abnormal state features are identified according to the feature fusion matrix.

5. The battery production defect inspection method according to claim 1, characterized in that: The multi-level screening of each of the key frame sequences based on the abnormal state characteristics to obtain a target key frame with defects includes: Establishing a hierarchical screening rule base according to the type of the abnormal state characteristics; Performing a preliminary screening on the key frame sequence based on the hierarchical screening rule base to obtain a first abnormal frame set; Performing adjacent frame correlation analysis on the first abnormal frame set to obtain a second abnormal frame set; Verify and screen the second-level abnormal frame set based on the process state transition rule; The keyframes that pass the validation filter are determined as target keyframes.

6. The battery production defect inspection method according to claim 1, characterized in that: The performing spatiotemporal mapping on the target key frame to determine the associated workstations corresponding to the abnormal production line includes: Constructing a topological structure diagram of the workstations corresponding to each of the battery production lines; Extracting process timestamps and spatial coordinates from the target keyframe; Based on the process timestamp and the spatial coordinates, the associated workstations corresponding to the abnormal production line are located in the workstation topology structure diagram.

7. The battery production defect inspection method according to claim 1, characterized in that: The step of performing a graded evaluation on the abnormal state characteristics according to the process parameters of the associated workstations and generating abnormal warning information includes: Obtaining the process parameters of the associated workstation and the corresponding process standard range; Calculating the deviation of the process parameters from the process standard range; Quantitatively scoring the abnormal state feature based on the deviation, and determining the abnormality level according to the quantitative score; The impact of the abnormality level on the battery production line is analyzed, and graded abnormality warning information including the abnormality level and the impact level is generated.

8. A battery production defect inspection system, characterized in that: The system comprises: A frame sequence extraction module, used to obtain the video stream corresponding to each battery production line, and extract the key frame sequence corresponding to each video stream; A feature analysis module is used to perform a fusion analysis of spatial features and temporal features on each of the key frame sequences to obtain abnormal state features; A defect detection module, used for performing multi-level screening on each of the key frame sequences based on the abnormal state characteristics to obtain target key frames with defects; An abnormality determination module, used to perform spatiotemporal mapping on the target key frame and determine the associated workstations corresponding to the abnormal production line; The abnormal warning module is used to perform graded evaluation on the abnormal state characteristics according to the process parameters of the associated workstations and generate abnormal warning information.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor and executing the method according to any one of claims 1 to 7.

10. An electronic device, characterized in that: It includes a processor, a memory, a user interface and a network interface, the memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device executes the method as described in any one of claims 1-7.

Images