Battery transfer detection method, device and equipment and storage medium

Abstract

The invention belongs to the technical field of logistics line detection, and discloses a battery transfer detection method, device and equipment and a storage medium, and the method comprises the steps: obtaining image data of a plurality of positions on a logistics line; performing target detection on the image data to obtain an object position and an object attribute; and battery transfer on the logistics line is detected according to the object position and the object attribute, a detection result is obtained, and when the detection result is that the batteries on the logistics line are transferred abnormally, battery abnormal transfer reminding is carried out. Target detection is carried out on the image data of the multiple positions on the logistics line, so that the batteries which are abnormally transferred are quickly identified, the transfer condition of the batteries on the logistics line can be effectively monitored, the abnormal transfer phenomenon can be found in time, and the safety and efficiency of logistics management are improved.

Description

Technical Field

[0001] The present application relates to the field of logistics line detection technology, and in particular to battery transport detection methods, devices, equipment and storage media. Background Art

[0002] In the lithium battery manufacturing process, there is a phenomenon of personnel inspecting materials in violation of regulations during the visual inspection process after ultrasonic welding in the battery cell process. The battery transportation process needs to be controlled by quality inspectors, and it is impossible to monitor in real time whether the rework process is standardized.

[0003] At present, the logistics line transfer inspection is mainly carried out by quality inspectors regularly checking whether employees have violated the regulations in material inspection. Real-time monitoring cannot be guaranteed and the inspection efficiency is low. Summary of the invention

[0004] The main purpose of this application is to provide a battery transport detection method, device, equipment and storage medium, aiming to solve the technical problem of low efficiency in abnormal object detection on the logistics line.

[0005] In a first aspect, the present application provides a battery transport detection method, the battery transport detection method comprising the following steps: Acquire image data at multiple locations along the logistics line; Perform target detection on image data to obtain the location and attributes of objects on the logistics line; The battery transportation on the logistics line is detected according to the object position and object attributes to obtain the detection results.

[0006] In this solution, by performing target detection on image data at multiple locations on the logistics line, we can quickly identify batteries with abnormal transportation, effectively monitor the transportation of batteries on the logistics line, discover abnormal transportation phenomena in time, and improve the safety and efficiency of logistics management.

[0007] In some embodiments, the step of detecting abnormal transport of objects on a logistics line according to the location and attributes of the objects and obtaining the detection result includes: Determine an object attribute as a target to be tracked having a first label; Determine a first object rectangular frame position of the target to be tracked based on the object position, and determine a first rectangular frame according to the first object rectangular frame position; Calculate a first overlapping ratio between the first rectangular frame and the preset area of ​​the logistics line; When the first overlap ratio is less than a first preset ratio threshold, adding the first object rectangular frame position to a to-be-tracked list; Based on the to-be-tracked list, a battery transfer detection is performed on the to-be-tracked target to obtain a detection result.

[0008] In the technical solution of the embodiment of the present application, by setting specific labels and overlap ratio thresholds, objects that need attention can be screened out more accurately, reducing the false alarm rate and improving the reliability of anomaly detection.

[0009] In some embodiments, the step of performing battery transport detection on the target to be tracked based on the to-be-tracked list and obtaining the detection result includes: Obtaining a first rectangular frame position of a target to be tracked in the to-be-tracked list, and obtaining a first rectangular frame according to the first rectangular frame position; Determine the state of the target to be tracked according to the first rectangular frame; When the state of the target to be tracked is a preset state, it is determined that the target to be tracked has abnormal transport, and a detection result of abnormal transport of the battery is obtained.

[0010] In the technical solution of the embodiment of the present application, the state of the target to be tracked is determined according to the first rectangular frame, so as to determine whether there is abnormal transportation of batteries on the logistics line through the state of the target to be tracked, thereby improving the accuracy and speed of abnormal battery transportation detection.

[0011] In some embodiments, the step of determining the state of the target to be tracked according to the first rectangular frame includes: When a first overlap ratio between a first rectangular frame of the target to be tracked and a preset area of ​​the logistics line is greater than a second preset ratio threshold and the object attribute of the target to be tracked is a first label, changing the object attribute of the target to be tracked to a second label; Obtain the initial position and current position of the target to be tracked according to the list to be tracked; Calculate a first distance between the initial position and the center position of the logistics line and a second distance between the current position and the center position of the logistics line; Determine the number of image frames of the target to be tracked in the to-be-tracked list according to the first rectangular frame; The state of the target to be tracked is determined according to the second tag, the first distance, the second distance, the number of image frames, and the first overlapping ratio.

[0012] In the technical solution of the embodiment of the present application, an object attribute change mechanism and distance calculation are added, which can more flexibly adapt to changes in the position of objects on the logistics line and improve the dynamic adaptability of anomaly detection.

[0013] In some embodiments, the step of determining the state of the target to be tracked according to the second tag, the first distance, the second distance, the number of image frames, and the first overlap ratio includes: Under the conditions that the first distance is less than the second distance, the object attribute of the target to be tracked is changed to the second label, the first overlap ratio is less than the third preset ratio threshold, and the number of image frames reaches the set number of frames, determining that the state of the target to be tracked is the preset state, the second preset ratio threshold is less than the third preset ratio threshold, and the first preset ratio threshold is greater than the third preset ratio threshold; If at least one of the following conditions is not satisfied: the first distance is less than the second distance, the object attribute of the target to be tracked is changed to the second label, the first overlap ratio is less than the third preset ratio threshold, and the number of image frames reaches the set number of frames, it is determined that the state of the target to be tracked is not the preset state.

[0014] In the technical solution of the embodiment of the present application, by setting a combination of multiple conditions to determine the state of the target to be tracked, abnormal transportation behavior of batteries on the logistics line can be identified more accurately, reducing the probability of false detection.

[0015] In some embodiments, the method further comprises: Predict the target to be tracked in the tracking list and obtain the predicted rectangular frame position of the object; Obtain a predicted rectangular frame according to the predicted object rectangular frame position; Calculate the overlap value between the predicted rectangular box and the corresponding first rectangular box in the list to be tracked; When the overlap value is greater than a second preset ratio threshold, determining that the predicted rectangular box matches the first rectangular box; When the predicted rectangular box matches the first rectangular box, the position of the first object rectangular box of the target to be tracked is updated according to the position of the predicted object rectangular box to obtain an updated list to be tracked.

[0016] In the technical solution of the embodiment of the present application, by predicting the future position of the object in advance and comparing the predicted position of the object with the current position, the movement state of the object on the logistics line can be quickly updated in real time, which helps to take measures in advance and reduce missed detections or false detections caused by sudden changes in the object.

[0017] In some embodiments, the method further comprises: When the overlap value is less than or equal to a second preset ratio threshold, determining that the predicted rectangular box does not match the first rectangular box; When the predicted rectangular frame does not match the first rectangular frame, obtaining the number of unmatched frames; When the number of unmatched frames is less than or equal to the preset frame number threshold, the target corresponding to the predicted rectangular box is added to the to-be-tracked list for detection; When the number of unmatched frames is greater than a preset frame number threshold, the target to be tracked corresponding to the predicted rectangular box is removed from the list to be tracked.

[0018] In the technical solution of the embodiment of the present application, when the predicted position does not match the actual position, a reasonable processing flow is set up, which can not only update the tracking list in time, but also avoid long-term erroneous tracking, thereby improving the accuracy of detection.

[0019] In some embodiments, the method further comprises: When the detection result shows that there is abnormal transportation of batteries on the logistics line, a reminder of abnormal battery transportation will be issued.

[0020] In the technical solution of the embodiment of the present application, when abnormal transportation of batteries on the logistics line is detected, an abnormal battery transportation reminder is issued, which can immediately attract the attention of relevant personnel, respond quickly to abnormal situations, reduce the occurrence of abnormal situations such as battery cells being taken out of the logistics line, and reduce losses.

[0021] In some embodiments, when the detection result shows that the battery on the logistics line is transported abnormally, the step of providing a reminder for abnormal battery transport includes: When the detection result shows that there is abnormal transportation of batteries on the logistics line, an alarm message is generated and the historical alarm interval time is obtained; When the historical alarm interval exceeds the preset interval, the center point data of the target to be tracked is obtained; Generate trajectory information of the target to be tracked based on the center point data; The alarm information and trajectory information are sent to the user to remind the abnormal battery transportation.

[0022] In the technical solution of the embodiment of the present application, after abnormal transportation is detected, an alarm message is generated and sent to the user, which can immediately attract the attention of relevant personnel, quickly respond to abnormal situations, reduce the occurrence of abnormal situations such as battery cells being taken out of the logistics line, and reduce losses.

[0023] In some embodiments, the battery transportation on the logistics line is detected according to the object position and object attributes, and the step of obtaining the detection result includes: Determine that the object attribute is a preset object of a third tag; Obtaining a second rectangular frame position of the preset object according to the object position, and determining a second rectangular frame according to the second rectangular frame position; Calculating a second overlapping ratio between the second rectangular frame and the preset area of ​​the logistics line; When the second overlap ratio is greater than a fourth preset ratio threshold, determining that a preset object exists on the logistics line; Obtain the number of detections of the presence of preset objects on the detection logistics line; When a preset object is detected within a preset time threshold and the number of detections is greater than a preset number threshold, it is determined that abnormal transportation of batteries on the logistics line exists, and a detection result of abnormal transportation of batteries is obtained.

[0024] In the technical solution of the embodiment of the present application, whether a preset object is detected is determined based on the label of the object, and abnormal battery transportation detection is performed based on whether the preset object exists on the logistics line, thereby improving the comprehensiveness and accuracy of the detection.

[0025] In some embodiments, the method further comprises: When the detection result shows that there is abnormal transportation of batteries on the logistics line, the preset object or the target to be tracked is marked to obtain a marked image; Generate alarm information and obtain historical alarm intervals; When the historical alarm interval exceeds the preset interval, the alarm information and the annotated image are sent to the user to remind the user of abnormal battery transportation.

[0026] In the technical solution of the embodiment of the present application, in the event of abnormal battery transportation, an alarm message is generated in a timely manner to provide a reminder. By generating an annotated image, not only can the user intuitively understand the abnormal situation, but it can also serve as an important basis for subsequent analysis.

[0027] In some embodiments, the battery transportation on the logistics line is detected according to the object position and object attributes, and the step of obtaining the detection result includes: Inputting object locations and object attributes into an anomaly detection model; The abnormal transportation of objects on the logistics line is detected through the anomaly detection model to obtain the detection results.

[0028] In the technical solution of the embodiment of the present application, by using a trained anomaly detection model for detection, complex logic can be encapsulated inside the model, which simplifies the external interface, facilitates maintenance and upgrading, and also improves the accuracy and speed of detection.

[0029] In some embodiments, the anomaly detection model is trained by: Acquire training samples, wherein the training samples include normal battery transport images and abnormal battery transport images; The initial model is trained according to the training samples to obtain the anomaly detection model.

[0030] In the technical solution of the embodiment of the present application, a model is trained using training samples containing normal transport images and abnormal transport images of batteries to obtain an abnormality detection model. The abnormality detection model can distinguish between normal and abnormal states more quickly and accurately, thereby improving the accuracy and efficiency of detection.

[0031] In some embodiments, obtaining a training sample includes: Get the original image data; Label the original image data to obtain the original object position and original object attributes on the logistics line; Objects on the logistics line are tracked according to the original object positions and original object attributes to obtain normal battery transportation images and abnormal battery transportation images.

[0032] In the technical solution of the embodiment of the present application, by annotating the original image data and clarifying the position and attributes of objects on the logistics line, it is possible to ensure that the data set used for training has high accuracy and high reliability. High-quality data is the basis for building an efficient model, which helps to improve the learning efficiency and final performance of the model. Based on the annotated original object position and attribute information, accurate tracking of objects on the logistics line can be achieved. This accurate tracking capability is crucial for distinguishing normal transportation from abnormal transportation, and helps the model learn image features under different transportation states, further improving the accuracy of anomaly detection.

[0033] In a second aspect, in order to achieve the above-mentioned purpose, the present application also proposes a battery transport detection device, the battery transport detection device comprising: An acquisition module, used to acquire image data of multiple locations on the logistics line; The detection module is used to detect targets on image data and obtain the location and attributes of objects on the logistics line; The detection module is also used to detect the battery transportation on the logistics line according to the object position and object attributes to obtain the detection result.

[0034] On the third aspect, in order to achieve the above-mentioned purpose, the present application also proposes a battery transport detection device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the battery transport detection method as described above.

[0035] Fourthly, in order to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by the processor, the steps of the battery transport detection method as described above are implemented.

[0036] In a fifth aspect, to achieve the above-mentioned purpose, the present application also provides a computer program product, the computer program product includes a computer program, and when the computer program is executed by a processor, the steps of the battery transport detection method as described above are implemented.

[0037] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present application. Moreover, the same figure numbers are used throughout the drawings to represent the same components. In the drawings: Figure 1 A schematic diagram of a flow chart of an embodiment of a battery transport detection method proposed in an embodiment of the present application; Figure 2 A schematic diagram of the relationship between the logistics line area, the center line, and the objects on the logistics line in an embodiment of the battery transport detection method proposed in the present application; Figure 3 A schematic diagram of a flow chart of an embodiment of a battery transport detection method proposed in an embodiment of the present application; Figure 4 A schematic diagram of a flow chart of an embodiment of a battery transport detection method proposed in an embodiment of the present application; Figure 5 A schematic diagram of a flow chart of an embodiment of a battery transport detection method proposed in an embodiment of the present application; Figure 6 A schematic diagram of a flow chart of an embodiment of a battery transport detection method proposed in an embodiment of the present application; Figure 7 A brief flowchart of an embodiment of a battery transport detection method proposed in an embodiment of the present application; Figure 8 This is a schematic diagram of the module structure of the battery transport detection device according to an embodiment of the present application; Fig. 9 Schematic diagram of the device structure of the hardware operating environment involved in the battery transport detection method in the embodiment of the present application.

[0039] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0040] The following embodiments of the technical solution of the present application are described in detail in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present application, and are therefore only used as examples, and cannot be used to limit the scope of protection of the present application.

[0041] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" in the specification and claims of this application and the above-mentioned figure descriptions and any variations thereof are intended to cover non-exclusive inclusions.

[0042] In the description of the embodiments of the present application, the technical terms "first", "second", etc. are only used to distinguish different objects, and cannot be understood as indicating or implying relative importance or implicitly indicating the number, specific order or primary and secondary relationship of the indicated technical features. In the description of the embodiments of the present application, the meaning of "multiple" is more than two, unless otherwise clearly and specifically defined.

[0043] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0044] In the description of the embodiments of the present application, the term "and / or" is only a description of the association relationship of the associated objects, indicating that there may be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0045] In the description of the embodiments of the present application, the term "multiple" refers to more than two (including two). Similarly, "multiple groups" refers to more than two groups (including two groups), and "multiple pieces" refers to more than two pieces (including two pieces).

[0046] In the description of the embodiments of the present application, the technical terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc. indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, which are only for the convenience of describing the embodiments of the present application and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the embodiments of the present application.

[0047] In the description of the embodiments of the present application, unless otherwise clearly specified and limited, technical terms such as "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, and it can be the internal connection of two elements or the interaction relationship between two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the embodiments of the present application can be understood according to the specific circumstances.

[0048] In the ultrasonic welding process of the cell process section in battery production, the pole ear and the adapter are ultrasonically welded. After the welding is completed, it is transferred to the next process through the logistics line. It is necessary to check whether the welding part is flat and has no obvious bumps to ensure that the welding quality meets the standards. In addition, it is crucial to confirm whether the welding joint is uniform and crack-free for the safety and reliability of the battery. Otherwise, the next process may have problems such as equipment downtime, logistics line jamming, and equipment inability to continue production. In order to increase production capacity, employees take out the cells on the logistics line (may be held, placed in a tray or on an acrylic board) and transfer them to the shelf. During the cell removal process, the cell opening angle may be too large, causing the pole ear to crack. It is necessary to investigate the cause and confirm the risk range, so it is necessary to stop the line and stop shipping, affecting production and shipping, and the loss cannot be estimated. Quality inspectors need to control the cell transfer process, but it is impossible to monitor the rework process in real time to ensure that it is standardized. Therefore, it is necessary to use AI-assisted detection to identify whether there are any illegal operations on the logistics line of the ultrasonic welding process (that is, when personnel transfer the cells from the logistics line to other areas).

[0049] At present, the main method for object detection in logistics lines is to track objects using DeepSort (Deep learning based SORT, a multi-target tracking algorithm based on deep learning). The deep learning model is used to extract the appearance features of the target, and the target's motion information and appearance features are combined for associated tracking. The matching accuracy is improved, but the efficiency is reduced. In addition, during the detection process, only the target holding the battery cell that shows signs of leaving the logistics line is tracked in combination with the actual situation on site. Therefore, the target filtering and cascade matching are only based on the target's motion position information, and the detection is not accurate enough.

[0050] Therefore, it is necessary to improve the efficiency and accuracy of abnormal transport detection of objects on the logistics line. The inventive concept of this application is: According to the actual target effect, a lightweight target tracking algorithm is used to add initial filtering conditions for the tracking target. According to the tracking results, the initial direction of the tracking trajectory and the intersection and union ratio of the tracking target and the logistics line area are calculated to accurately locate whether the battery cell is taken out of the logistics line abnormally. This improves the efficiency and accuracy of abnormal object transportation detection.

[0051] Specifically, this solution performs target detection on the image data acquired on the logistics line to identify the positions and attributes of multiple objects, and continuously tracks the objects based on the positions and attributes of each object. According to the tracking results, it is determined whether there is abnormal transportation of battery cells on the logistics line, that is, whether the user has violated any regulations.

[0052] In practical applications, the embodiments described in this application are applied to the post-ultrasonic welding process of the battery cell process in the battery production process, and can also be applied to other processes in the battery production process, which is not limited by this embodiment.

[0053] This application proposes a battery transport detection method to solve the technical problem of low efficiency of abnormal object detection on logistics lines. Figure 1 In this example, the battery transport detection method includes: Step S10: Acquire image data of multiple locations on the logistics line.

[0054] It should be noted that the executor of this embodiment is a battery transportation detection device, which may specifically include: a camera for collecting online logistics videos, a processor for image processing and target detection, and equipment for early warning, etc.

[0055] In the specific implementation, the logistics line is mainly an ultrasonic logistics line, and multiple positions can be set in advance, specifically various positions that can be set in the ultrasonic logistics line area, such as the positions around the logistics line area that can cover the entire logistics line area. Multiple positions are positions for shooting ultrasonic logistics lines. In the specific implementation, multiple cameras can be installed in advance at various positions on the ultrasonic logistics line for on-site monitoring. For example, cameras can be installed at the entrance, exit, key nodes, etc. of the logistics line to ensure coverage of the entire logistics line.

[0056] It is understandable that the image data can be obtained by segmenting the video data taken by the camera. Specifically, the video streams of surveillance cameras at multiple locations on site can be obtained through RTSP (Real-Time Streaming Protocol, network control protocol) to obtain video data, and frame extraction can be performed based on the video data to obtain image data.

[0057] For example, if the real-time video stream has 25 frames per second, an image is taken every 5 frames of each video stream to obtain image data at multiple locations.

[0058] Step S20: Perform target detection on the image data to obtain the object position and object attributes.

[0059] In a specific implementation, target detection on image data can be achieved through deep learning target detection algorithms, such as R-CNN (Region-based Convolutional Neural Network), SPP-Net (Spatial Pyramid Pooling Networks), Fast R-CNN (Fast Region-based Convolutional Network), etc. It can also be achieved through target detection algorithms, such as corner-based detection algorithms, which determine the bounding box by predicting the upper left and lower right corners of the target; target detection can also be achieved through target detection models, such as YOLO-v5, YOLOv10 and other series of target detection models.

[0060] For example, the target detection model trained using the YOLO-v5 model is used to process the acquired image data to identify the location and attributes of objects in the image.

[0061] The object location may include the specific location coordinates of the object on the logistics line, and the object attributes may include the category to which each object belongs. For example, label information can be used to represent the categories of different objects, such as battery cells on the logistics line (label 1), people holding battery cells (label 0), acrylic plates (label 2), etc.

[0062] The target detection model can be trained by the sample data collected in the early stage with the object position and object attributes annotated, so as to identify the target of the image data through the target detection model and obtain the object position and object attributes on the current logistics line. The objects may include battery cells, people holding battery cells, acrylic plates, etc., and may also include other objects.

[0063] Before detecting abnormal transfer of objects on the logistics line, data can be collected in advance and point information can be configured. Specifically, RTSP can be used to obtain the video of the on-site surveillance camera, collect the video data of normal and abnormal operation behaviors of the corresponding points of the ultrasonic logistics line, divide the video into image data, select images with large screen changes and cover various situations as much as possible, and process the data. The specific steps are as follows: Video collection: Simulate various cell removal actions according to the actual rework process that may occur, and intercept the corresponding video, including the following situations: OK (no abnormal transfer): employees wear white / blue / black gloves to transfer the cells alone / with trays from outside the logistics line to inside the logistics line; employees wear white / blue / black gloves to check the cells normally, and the cells do not completely exceed the logistics line. NG (abnormal transfer): employees wear white / blue / black gloves to transfer the cells alone / with trays from inside the logistics line to outside the logistics line; white / pink acrylic plates appear on the cover of the logistics line for more than a certain time (user-specified time). Both of the above situations need to be simulated with battery cells of as many appearances as possible, including but not limited to: half blue film and half white film battery cells, battery cells fully covered with blue film, battery cells with a small middle part covered with blue film, etc.

[0064] Set business configuration information: The information such as the quadrilateral region, the center straight line steel_line, and the maximum existence time of the acrylic plate that determines the logistics line at each location is exposed to the user through the interface and configured according to the actual screen changes. Figure 2 As shown, Figure 2 This is a schematic diagram of the relationship between the logistics line area, the center line, and objects on the logistics line. The battery cells and pallets form a whole and are placed on the logistics line. Users can check the battery cells by picking up the pallet.

[0065] After setting the business configuration information, the collected data can be processed to build a model and obtain a YOLO-v5 model. Specifically, the collected video can be divided into frames and the images with high screen repetition can be filtered out from the image data as the sample image data set, which can be divided into a sample training set, a sample verification set, and a sample test set in proportion. Similarity calculation, repetition rate comparison, and other methods can be used to filter out images with high repetition, so as to eliminate images with high repetition. After obtaining the sample training set, sample verification set, and sample test set, the sample training set and sample verification set of the image data can be annotated by annotation software (such as labelme annotation software) or manually, annotating the attributes and positions of objects such as the employee holding the battery cell (label 0), a separate battery cell (label 1), and an acrylic plate (label 2), and trained by the yolov5s model to obtain a target detection model. The trained target detection model is used to pre-annotate the sample test set data (for example, using annotation software for annotation), and the annotation files with incorrect labels or annotation box sizes that do not meet expectations in the pre-annotated files are modified, and the model training iteration is performed. Thus, an updated target detection model is obtained.

[0066] When abnormal transport detection is required on a logistics line, the acquired image data can be input into the target detection model for detection, thereby obtaining the object position and object attributes of each object. The object position may include the position of the object rectangular frame. In addition to outputting the object position and object attributes, the target detection model can also output the confidence of the detected object, so that the category of the object can be identified based on the confidence.

[0067] Step S30: Detect the battery transportation on the logistics line according to the object position and object attributes to obtain the detection result, wherein the battery transportation on the logistics line includes: tracking the object according to the object position and object attributes to determine whether there is abnormal battery transportation.

[0068] In a specific implementation, the overlap ratio between each object rectangular frame and the preset area can be calculated according to the object attributes and the object position, so that the object can be further tracked and detected according to the overlap ratio, so as to detect whether there is abnormal transfer of batteries on the logistics line, that is, the batteries on the logistics line are taken out from the logistics line by the user.

[0069] For example, the preset area of ​​the logistics line is set as a quadrilateral area, and the intersection over union (IoU) of each object's rectangular box and the area is calculated. If the IoU of an object is less than the preset threshold (such as 0.9), it is considered that the object may have been removed from the logistics line, and the object needs to be continuously tracked and a comparison rule is set, so that the tracking result is compared with the set comparison rule. If the comparison rule is not met, it means that the battery has been transported abnormally.

[0070] This embodiment proposes a battery transfer detection method, which performs target detection on image data of multiple locations on the logistics line, thereby quickly identifying batteries with abnormal transfer. It can effectively monitor the transfer status of batteries on the logistics line, discover abnormal transfer phenomena in time, and improve the safety and efficiency of logistics management.

[0071] In some embodiments, the object can be tracked and detected according to the specific rectangular frame position in the object position, so as to obtain the object detection result. Figure 3 , step S30 specifically includes: Step S301: Determine an object to be tracked with a first tag as an attribute of the object.

[0072] It should be noted that the first tag, that is, the tag corresponding to the person holding the battery cell (tag 0), can be used to filter out the target with the first tag from the tags of each object in the object attributes as the target to be tracked.

[0073] When employees check batteries normally, the target holding the batteries will remain within the logistics line and does not need to be tracked. Therefore, the target to be tracked of the handheld batteries can be determined first, and the target to be tracked can be tested to determine whether continuous tracking is required.

[0074] Step S302: determining a first object rectangular frame position of the target to be tracked based on the object position, and determining a first rectangular frame according to the first object rectangular frame position.

[0075] It should be noted that after the target detection is completed, the specific position information of each object rectangular frame can be obtained according to the object position, including the coordinates of the upper left corner and the lower right corner of the rectangular frame.

[0076] By obtaining the position of the object's rectangular frame, it is convenient to continuously track and detect the object based on the position of the object's rectangular frame.

[0077] It can be understood that the first object rectangular frame position is the rectangular frame position of the handheld battery cell to be tracked. After the object position is obtained, the first rectangular frame can be determined, including the size and attributes of the first rectangular frame.

[0078] There may be multiple persons holding batteries, and therefore, there may be multiple targets to be tracked, and the number of obtained first rectangular frames may also be multiple.

[0079] When the object attributes are different, the corresponding detection objects are different. For example, if the object attribute is 1, that is, the object is a separate battery cell, there is no need to perform abnormal transfer detection. If the object attribute is 0, that is, the employee is holding the battery cell, the target with the object attribute of 0 can be continuously detected according to the position of the object rectangular frame.

[0080] For example, if the overlap ratio between the rectangular frame position of a person holding a battery cell (label 0) and the preset area of ​​the logistics line is less than 0.9, it is considered that the person may move the battery cell out of the logistics line, so further detection of the object with the label is required.

[0081] Step S303: Calculate a first overlapping ratio between the first rectangular frame and the preset area of ​​the logistics line.

[0082] In a specific implementation, the overlap ratio between each rectangular frame in the first rectangular frame and the preset area of ​​the logistics line, ie, the first overlap ratio, can be calculated respectively. The preset area of ​​the logistics line is the quadrilateral area of ​​the logistics line.

[0083] Calculating the overlap ratio between the first rectangular frame and the quadrilateral area of ​​the logistics line can calculate the intersection area and union area between the first rectangular frame and the preset area of ​​the logistics line, thereby obtaining the intersection and union ratio between the first rectangular frame and the preset area of ​​the logistics line, that is, the first overlap ratio. The calculation process is as follows: (Formula 1) In the above formula 1, Area of ​​Overlap is the area of ​​the intersection area between the first rectangular frame and the rectangular frame formed by the preset area of ​​the logistics line. Area of ​​Union is the area of ​​the union area between the first rectangular frame and the rectangular frame formed by the preset area of ​​the logistics line.

[0084] Step S304: when the first overlap ratio is less than a first preset ratio threshold, the first object rectangular frame position is added to a to-be-tracked list.

[0085] The first preset ratio threshold can be set to 0.9, which is a critical value for measuring whether the target to be tracked needs to be detected. For example, if the first overlap ratio is greater than or equal to 0.9, it proves that the target corresponding to the first rectangular box remains within the logistics line and does not need to be tracked. When the first overlap ratio is less than the first preset ratio threshold 0.9, the position of the first rectangular box can be added to the list to be tracked.

[0086] The to-be-tracked list is for targets that may have abnormal transfers, so continuous tracking is required. When a new target appears, the corresponding target information can be added to the to-be-tracked list. When a target in the to-be-tracked list has not been tracked for a long time, it can also be removed from the to-be-tracked list.

[0087] Step S305: Perform battery transfer detection on the target to be tracked based on the list to be tracked to obtain a detection result.

[0088] By continuously tracking the targets in the tracking list and calculating the overlap ratio between the subsequent rectangular frame and the preset area of ​​the logistics line, it can be determined whether the battery has been removed from the logistics line.

[0089] If the target to be tracked is removed from the logistics line, that is, there is abnormal battery transportation, an alarm can be issued.

[0090] In the technical solution of the embodiment of the present application, by setting specific labels and overlap ratio thresholds, objects that need attention can be screened out more accurately, reducing the false alarm rate and improving the reliability of anomaly detection.

[0091] In a feasible implementation manner, whether the target to be tracked has abnormal transport can be further confirmed according to the status of each target to be tracked in the list to be tracked, so as to obtain the detection result, and step S305 includes: Step A11: obtaining the first rectangular frame position of the target to be tracked in the to-be-tracked list, and obtaining the first rectangular frame according to the first rectangular frame position; It should be noted that the rectangular frame position of each target to be tracked in the to-be-tracked list and the size of the first rectangular frame may be obtained, thereby facilitating determination of the state of the target to be tracked.

[0092] Step A12: determining the state of the target to be tracked according to the first rectangular frame; The state of the target to be tracked may be determined based on the information of the first rectangular frame. For example, the state of the target to be tracked may be determined based on the change in the first overlapping ratio between the first rectangular frame and the preset area of ​​the logistics line.

[0093] The state of the target to be tracked may be a preset state or a non-preset state. The preset state indicates an NG state of the target to be tracked, that is, a state in which the target to be tracked is in abnormal transport.

[0094] In a feasible implementation, determining the state of the target to be tracked based on the first rectangular frame includes: when the first overlap ratio between the first rectangular frame of the target to be tracked and the preset area of ​​the logistics line is greater than the second preset ratio threshold and the object attribute of the target to be tracked is the first label, changing the object attribute of the target to be tracked to the second label; obtaining the initial position and the current position of the target to be tracked based on the list to be tracked; calculating the first distance between the initial position and the center position of the logistics line and the second distance between the current position and the center position of the logistics line; determining the number of image frames of the target to be tracked in the list to be tracked based on the first rectangular frame; and determining the state of the target to be tracked based on the second label, the first distance, the second distance, the number of image frames and the first overlap ratio.

[0095] It should be noted that the second preset ratio threshold can be set to 0.1, that is, the intersection ratio between the first rectangular box of the target to be tracked and the preset area of ​​the logistics line is greater than 0.1, which proves that the target to be tracked has not left the logistics line at this time, and the object attribute of the target to be tracked is the first label 0. At this time, the object attribute of the target to be tracked is changed from the first label to the second label 1, that is, a separate battery cell appears.

[0096] Since there are multiple identical targets to be tracked or multiple different targets to be tracked at different times in the list to be tracked, the initial position and current position of the same target to be tracked can be queried in the list to be tracked. The initial position can be the position at the beginning of tracking the target to be tracked (that is, when the first rectangular frame of the target to be tracked and the preset area of ​​the logistics line are less than 0.9), and the current position is the current position of the target to be tracked. The center position of the logistics line can be any point on the center line of the logistics line. A fixed center position of the logistics line can be set on the center line of the logistics line in advance, so as to calculate the first distance between the initial position and the center position of the logistics line and the second distance between the current position and the center position of the logistics line. The first distance and the second distance can be the straight-line distance between the center position of the logistics line.

[0097] The number of image frames of the target to be tracked is the number of each target to be tracked added to the list to be tracked. To improve the tracking accuracy, the number of image frames of the target to be tracked added to the list to be tracked can be set, for example, set to 4 frames or 5 frames, thereby improving the detection accuracy.

[0098] In a specific implementation, whether the target to be tracked is in a preset state can be determined based on the second tag, the first distance, the second distance, the number of image frames, and the first overlapping ratio.

[0099] By adding object attribute change mechanisms and distance calculations, it can more flexibly adapt to changes in the position of objects on the logistics line, improving the dynamic adaptability of anomaly detection.

[0100] In a feasible implementation manner, the step of determining the state of the target to be tracked according to the second tag, the first distance, the second distance, the number of image frames, and the first overlap ratio includes: Under the conditions that the first distance is less than the second distance, the object attribute of the target to be tracked is changed to the second label, the first overlap ratio is less than the third preset ratio threshold, and the number of image frames reaches the set number of frames, determining that the state of the target to be tracked is the preset state, the second preset ratio threshold is less than the third preset ratio threshold, and the first preset ratio threshold is greater than the third preset ratio threshold; If at least one of the following conditions is not satisfied: the first distance is less than the second distance, the object attribute of the target to be tracked is changed to the second label, the first overlap ratio is less than the third preset ratio threshold, and the number of image frames reaches the set number of frames, it is determined that the state of the target to be tracked is not the preset state.

[0101] It should be noted that the third preset ratio threshold can be set to 0.2. If the first overlap ratio is less than the third preset ratio threshold, it means that the target to be tracked is gradually leaving the logistics line but has not completely left the logistics line. The first distance is less than the second distance, which means that the initial position of the target to be tracked is closer to the center straight line of the logistics line than it is now. The object attribute of the target to be tracked is changed to the second label, which means that the target to be tracked has become a separate battery cell. The number of image frames of the target to be tracked reaches the set number of frames, which means that the tracking amount of the target to be tracked meets the requirements. The set number of frames can be set to 4 frames, 6 frames, etc.

[0102] When the target to be tracked meets the above four conditions at the same time, it means that there is an illegal operation by a person at this time, that is, someone takes the battery cell on the logistics line out of the logistics line, that is, the state of the target to be tracked is the preset state.

[0103] It is understandable that if at least one of the following conditions is not met: the tracking times of the target to be tracked reaches the set number of frames, the object attribute of the target to be tracked is not changed to the second label, the first overlap ratio of the target to be tracked and the preset area of ​​the logistics line is less than 0.2, or the initial position of the target to be tracked is closer to the center straight line than it is now, it means that there is no illegal operation of removing the battery cell from the logistics line, the current state of the target to be tracked is not the preset state, and the target to be tracked can be continuously tracked.

[0104] By setting a combination of multiple conditions to determine the status of the target to be tracked, abnormal battery transportation behavior can be identified more accurately and the probability of false detection can be reduced.

[0105] Step A13: When the state of the target to be tracked is a preset state, it is determined that the target to be tracked has abnormal transport, and a detection result of abnormal transport of the battery is obtained.

[0106] In a specific implementation, if the state of the target to be tracked meets the preset state, it means that the battery on the logistics line is taken out from the logistics line by the user, and the battery is abnormally transported.

[0107] In the technical solution of the embodiment of the present application, the state of the target to be tracked is determined according to the first rectangular frame, so as to determine whether abnormal transport exists according to the state of the object, thereby improving the accuracy and speed of abnormal transport detection.

[0108] It should be noted that in the process of tracking the target, there may be new targets appearing or disappearing, so the list to be tracked needs to be updated in real time. Figure 4 , after step S304, further comprising: Step A141: predict the target to be tracked in the tracking list to obtain the predicted rectangular frame position of the object.

[0109] In a specific implementation, the prediction of the target to be tracked in the tracking list can be achieved using Kalman filtering, a physical model-based method (such as a dynamic model or a kinematic model), a motion pattern-based method or a deep learning method. For example, the Kalman filter is used to predict the transport target, so as to obtain the next state of the target, that is, to predict the position of the rectangular box of the object.

[0110] Step A142: Obtain a predicted rectangular frame according to the predicted position of the object rectangular frame.

[0111] In a specific implementation, after the predicted object rectangular frame position is obtained, predicted data including the predicted rectangular frame may be obtained according to the predicted object rectangular frame position.

[0112] The prediction rectangle includes the ID of the target to be tracked, the category of the target to be tracked, etc.

[0113] Step A143: Calculate the overlap value between the predicted rectangular box and the corresponding first rectangular box in the to-be-tracked list.

[0114] In a specific implementation, the degree of overlap between the first rectangular frame and its corresponding predicted rectangular frame may be calculated to obtain an overlap value.

[0115] Step A144: when the overlap value is greater than the second preset ratio threshold, determine that the predicted rectangular box matches the first rectangular box.

[0116] It should be understood that the matching relationship between the calculated overlap value pair prediction rectangular box and the first rectangular block can be used to determine which predicted targets match the existing targets to be tracked, which are newly appeared targets, and which are disappeared targets.

[0117] Specifically, the second preset ratio threshold may be set to 0.1. When the overlap value is greater than the second preset ratio threshold 0.1, it may be determined that the predicted rectangular box matches the first rectangular box successfully.

[0118] In a feasible implementation manner, if the overlap value is less than or equal to the second preset ratio threshold, it is proved that the predicted rectangular box does not match the first rectangular box, and the unmatched rectangular box can be processed accordingly. Then, after step A143, the following is further included: When the overlap value is less than or equal to a second preset ratio threshold, determining that the predicted rectangular box does not match the first rectangular box; When the predicted rectangular frame does not match the first rectangular frame, obtaining the number of unmatched frames; When the number of unmatched frames is less than or equal to the preset frame number threshold, the target corresponding to the predicted rectangular box is added to the to-be-tracked list for detection; When the number of unmatched frames is greater than a preset frame number threshold, the target to be tracked corresponding to the predicted rectangular box is removed from the list to be tracked.

[0119] It should be noted that if the overlap value between the predicted rectangular box and the first rectangular box is less than or equal to the second preset ratio threshold, it proves that the predicted rectangular box does not match the first rectangular box. For the unmatched predicted rectangular box, it can be determined whether it exceeds the set number of frames max_age. Max_age is the maximum number of frames that survive in the tracking list when there is no target association. Therefore, the number of times the predicted rectangular box does not match the first rectangular box can be obtained to obtain the number of unmatched frames.

[0120] In a specific implementation, the preset frame number threshold, i.e., max_age, may be set to 5 frames, or may be set to other values, which is not limited in this embodiment.

[0121] If the number of unmatched frames is less than or equal to the preset frame number threshold, it is determined that the target trajectory is blocked, so the target corresponding to the predicted rectangular box can be added to the list to be tracked, and the update and detection can continue.

[0122] If the number of unmatched frames exceeds the preset frame number threshold, there is no need to track the target corresponding to the predicted rectangular box, and the target to be tracked corresponding to the predicted rectangular box can be removed from the list to be tracked, and the track of the target to be tracked can be deleted. Thus, the list to be tracked is updated according to the predicted rectangular box.

[0123] When the predicted location does not match the actual location, setting a reasonable processing flow can not only update the tracking list in time, but also avoid long-term erroneous tracking and improve the accuracy of detection.

[0124] Step A145: when the predicted rectangular frame matches the first rectangular frame, the position of the first object rectangular frame of the target to be tracked is updated according to the position of the predicted object rectangular frame to obtain an updated list to be tracked.

[0125] In a specific implementation, if the predicted rectangular frame matches the first rectangular frame, the position of the corresponding target in the target to be tracked can be updated according to the position of the predicted rectangular frame, and the position of the target to be tracked is continuously tracked, thereby performing battery transport detection.

[0126] It is understandable that after the matching is completed, the bounding boxes and IDs of all the targets to be tracked in the current frame can be returned, and the results can be added to the list to be tracked to obtain an updated list to be tracked. Therefore, when abnormal transport detection is performed on the target to be tracked, real-time detection can be performed through the updated list to be tracked.

[0127] In the technical solution of the embodiment of the present application, by introducing a prediction mechanism, the future position of an object can be predicted in advance, so that the predicted position of the object can be compared with the current position, and the movement state of the object on the logistics line can be quickly updated in real time, which helps to take measures in advance and reduce missed detections or false detections caused by sudden changes in the object.

[0128] In some embodiments, in addition to abnormal detection of battery cells, abnormal transport detection of objects also includes detecting whether there are objects carrying battery cells on the logistics line. Figure 5 , step S30 further includes: Step S301 ′: determining that the object attribute is a preset object of a third tag.

[0129] It should be noted that the object attribute is the third tag indicating the presence of a preset object on the logistics line. The preset object is a device that can carry the battery cells on the logistics line, such as an acrylic plate, a sealed box, etc. This embodiment does not limit this. This embodiment takes an acrylic plate as an example for illustration.

[0130] Step S302 ′: obtaining a second rectangular frame position of the preset object according to the object position, and determining a second rectangular frame according to the second rectangular frame position.

[0131] In a specific implementation, the position of the preset object rectangle frame, ie, the second rectangle frame position, can be obtained through the detected object rectangle frame position, so that the size, shape, etc. of the second rectangle frame can be determined according to the second rectangle frame position.

[0132] Step S303 ′: Calculate a second overlapping ratio between the second rectangular frame and the preset area of ​​the logistics line.

[0133] In a specific implementation, the overlap ratio between the second rectangular frame and the preset area of ​​the logistics line, that is, the second overlap ratio, can be calculated, and the second overlap ratio can be compared with the set ratio threshold to determine whether there is an acrylic plate on the logistics line.

[0134] In a specific implementation, when performing target detection, the confidence of the second rectangular box will also be output, so as to determine whether the detected target is a preset object based on the confidence. For example, if the confidence is greater than or equal to the set confidence threshold, the corresponding target is the preset object of the third label. If the confidence is less than the set confidence threshold, the corresponding target is not the preset object of the third label, and the rectangular box corresponding to the target can be removed from the detected objects.

[0135] Step S304 ′: when the second overlap ratio is greater than a fourth preset ratio threshold, determining that a preset object exists on the logistics line.

[0136] In a specific implementation, the fourth preset ratio threshold can be set to 0.5, or can be set to other values. For example, if the second overlap ratio is less than or equal to 0.5, it means that there is no preset object in the logistics line and there is no need to continue tracking it.

[0137] If the second overlap ratio is greater than 0.5, it means that the acrylic plate appears in the logistics line, and it is recorded as the acrylic plate is detected in the current frame, and the number of times the acrylic plate appears + 1 is used as the number of detections.

[0138] Step S305 ′: obtaining the number of detections of the presence of a preset object on the logistics line.

[0139] In a specific implementation, each time the acrylic plate is detected, the number of times the acrylic plate appears is increased by 1, so the number of detection times of the presence of the preset object can be obtained in real time.

[0140] Step S306': when the preset object is detected within the preset time threshold and the number of detections is greater than the preset number threshold, it is determined that there is abnormal transportation of the battery on the logistics line, and a detection result of abnormal transportation of the battery is obtained.

[0141] It should be understood that a preset number threshold and a preset time threshold can be set. The preset time threshold is the longest time that the acrylic plate can exist, plate_ng_thres. For example, if the preset time threshold is set to 5s, then every 5s, it is determined whether the current frame detects the acrylic plate and the number of detections during this period. If the current frame detects the acrylic plate and the number of detections is greater than the preset number threshold, it is determined that there is abnormal transportation of objects on the logistics line.

[0142] The preset number threshold is set to the preset time threshold. For example, if the preset time threshold is 5s, the preset number threshold is set to 5s. If the detection number is greater than 5s and the acrylic plate is detected in the current frame, it is recorded that the acrylic plate exists in the logistics line for a long time, proving that the user loads the battery cell through the acrylic plate, that is, it is determined that there is abnormal transportation of objects on the logistics line.

[0143] In the technical solution of the embodiment of the present application, whether a preset object is detected is determined based on the label of the object, and abnormal transportation detection is performed based on whether the preset object exists on the logistics line, thereby improving the comprehensiveness and accuracy of the detection.

[0144] In some embodiments, when the detection result of the object is that there is abnormal battery transportation, an alarm reminder can be issued in time, thereby reducing the occurrence of abnormal situations such as the battery cell being taken out of the logistics line and reducing losses. Figure 6 , after step S30, further comprising: Step S40: When the detection result shows that there is abnormal transportation of batteries on the logistics line, an abnormal transportation reminder of the batteries is issued.

[0145] It is understandable that if abnormal transport of the battery is detected, an alarm message can be generated and relevant information can be sent to the user. For example, an audible and visual alarm can be used to send an alarm on site, and the alarm message and image data of the abnormal transport can be pushed to the mobile phone or computer of the relevant personnel, so as to provide an abnormal transport reminder.

[0146] It is understandable that if abnormal transport of the battery is detected, an alarm message can be generated and relevant information can be sent to the user. For example, an audible and visual alarm can be used to send an alarm on site, and the alarm message and image data of the abnormal transport can be pushed to the mobile phone or computer of the relevant personnel, so as to provide an abnormal transport reminder.

[0147] In a feasible implementation, step S40 may include: Step B11: When the detection result shows that there is abnormal transportation of objects on the logistics line, an alarm message is generated and the historical alarm interval time is obtained.

[0148] It should be noted that if the detection result shows that there is abnormal transportation of batteries on the logistics line, an alarm message is generated and the historical alarm interval time is obtained. The historical alarm interval time is the interval time from the last alarm.

[0149] Step B12: When the historical alarm interval exceeds the preset interval, the center point data of the target to be tracked is obtained.

[0150] To avoid frequent alarms, subsequent alarm operations can be performed only when the historical alarm interval is greater than the set interval threshold. To improve the precise positioning of abnormal transported objects, the center point of the target to be tracked at different locations can be obtained to obtain the center point data.

[0151] Step B13: Generate trajectory information of the target to be tracked based on the center point data.

[0152] It should be noted that the trajectory line of the target to be tracked can be generated through the central point data of the target to be tracked, thereby obtaining the motion trajectory information of the target to be tracked.

[0153] Step B14: Send the alarm information and trajectory information to the user to remind the user of abnormal battery transportation.

[0154] It should be understood that after the trajectory information of the target to be tracked is generated, the alarm information and the trajectory information can be sent to the user together to provide an abnormal transfer reminder.

[0155] After obtaining the results of abnormal transport detection, the result data can be summarized and written into the database. At the same time, the result display data of abnormal transport can be mounted and written into the shared path through cifs (network file sharing protocol), and then pushed to the corresponding early warning platform and on-site alarm for alarm. According to the trajectory information, the entire violation process of the object with abnormal transport is saved to the corresponding path for archiving.

[0156] In the technical solution of the embodiment of the present application, after abnormal transportation is detected, an alarm message is generated and sent to the user, which can immediately attract the attention of relevant personnel, quickly respond to abnormal situations, reduce the occurrence of abnormal situations such as battery cells being taken out of the logistics line, and reduce losses.

[0157] In some embodiments, when the detection result of the object is that there is abnormal transportation, an alarm reminder can be issued in time, thereby reducing the occurrence of abnormal situations such as the battery cell being taken out of the logistics line and reducing losses. Step S40 may also include: Step C11: When the detection result shows that there is abnormal transfer of objects on the logistics line, the preset objects or the targets to be tracked are marked to obtain a marked image.

[0158] It should be noted that when the detection result shows that there is abnormal transfer of objects on the logistics line, the preset objects or the target to be tracked can be marked to obtain a marked image. The marked image is an image with violations.

[0159] Step C12: Generate alarm information and obtain historical alarm interval time.

[0160] When there is abnormal transfer, an alarm message is generated and the time of the last alarm is obtained.

[0161] Step C13: When the historical alarm interval exceeds the preset interval, the alarm information and the annotated image are sent to the user to remind the user of abnormal transfer.

[0162] When the time of the last alarm is greater than the preset interval time, the alarm information and the annotated image can be sent to the user to remind the abnormal transfer. For example, the annotated image can be pushed to the abnormal result feedback platform, that is, the network where the relevant person in charge is located, to remind the user and make an on-site alarm at the same time.

[0163] In the technical solution of the embodiment of the present application, when there is abnormal transportation, alarm information is generated in time to remind, and by generating an annotated image, not only can the user intuitively understand the abnormal situation, but it can also serve as an important basis for subsequent analysis.

[0164] In a feasible implementation, in order to improve the efficiency of abnormal transport detection, normal and illegal operation videos of the battery cell segment after ultrasonic welding can be used for model testing, so as to train the abnormal transport model and obtain a preset abnormal detection model. Therefore, step S30 may also include: Inputting object locations and object attributes into an anomaly detection model; The abnormal transportation of batteries on the logistics line is detected through the anomaly detection model to obtain the detection results.

[0165] It should be understood that the anomaly detection model is a model generated by training and testing the model through normal and illegal operation videos of the position after ultrasonic welding of battery cell segments, and after the model reaches the expected accuracy.

[0166] When performing abnormal transport detection, the object position and object attributes are input into the anomaly detection model, so that the abnormal transport of objects on the logistics line is detected through the anomaly detection model and the detection results are output.

[0167] In one feasible implementation, the training process of the anomaly detection model includes: Acquire training samples, wherein the training samples include normal battery transport images and abnormal battery transport images; The initial model is trained according to the training samples to obtain the anomaly detection model.

[0168] It should be noted that the training samples include normal battery transportation images and abnormal battery transportation images. The normal battery transportation images are images in which the batteries on the logistics line have not been taken out by the users, and the abnormal battery transportation images are images in which the batteries on the logistics line have been taken out by the users in violation of regulations.

[0169] After obtaining the normal battery transportation images and the abnormal battery transportation images, the initial model can be trained according to the normal battery transportation images and the abnormal battery transportation images. The initial model can be a neural network model or other types of models, and this embodiment does not limit this. By training the initial model, a model that can detect the transportation of batteries on the logistics line, that is, an abnormality detection model, is obtained.

[0170] By training the model using training samples containing normal transport images and abnormal transport images of batteries, an anomaly detection model is obtained. The anomaly detection model can distinguish between normal and abnormal states more quickly and accurately, thereby improving the accuracy and efficiency of detection.

[0171] In a feasible implementation, the normal battery transportation image and the abnormal battery transportation image can be annotated in advance, so the step of obtaining training samples includes: obtaining original image data; annotating the original image data to obtain the original object position and original object attributes on the logistics line; tracking the objects on the logistics line according to the original object position and the original object attributes to obtain the normal battery transportation image and the abnormal battery transportation image.

[0172] It should be noted that the original image data is an image containing the logistics line area collected by cameras arranged at various positions around the logistics line. The original image data can be annotated by labeling software (such as labelme labeling software) or manually to obtain the original object position and original object attributes on the logistics line in each image data. By tracking the objects on the logistics line based on the original object position and original object attributes, it can be determined whether there is abnormal battery transportation, and the normal battery transportation image and the abnormal battery transportation image can be obtained.

[0173] By annotating the original image data and clarifying the location and attributes of objects on the logistics line, we can ensure that the data set used for training has high accuracy and reliability. High-quality data is the basis for building an efficient model, which helps to improve the learning efficiency and final performance of the model. Based on the annotated original object location and attribute information, accurate tracking of objects on the logistics line can be achieved. This accurate tracking capability is crucial to distinguish between normal and abnormal transportation, and helps the model learn the image features under different transportation states, further improving the accuracy of anomaly detection.

[0174] In the technical solution of the embodiment of the present application, by using a trained anomaly detection model for detection, complex logic can be encapsulated inside the model, which simplifies the external interface, facilitates maintenance and upgrading, and also improves the accuracy and speed of detection.

[0175] For example, in order to help understand the implementation process of the battery transport detection method obtained by combining this embodiment with the above-mentioned embodiment 1, please refer to Figure 7 , Figure 7A brief flow chart of a battery transport detection method is provided, specifically: obtaining relevant configuration information and starting the detection process, reading the monitoring video stream of the point to be detected on the logistics line, judging whether the tracking target chain is empty, if so, creating a detection target link, and returning to the step of reading the monitoring video stream of the point to be detected, if the tracking target chain is not empty, performing target detection, detecting whether the intersection ratio of the hand-held battery cell target and the logistics line range is less than 0.9, if so, predicting the detection target through the Kalman filter, and using the Hungarian algorithm to match the tracking target, if not matched, deleting the tracking target, if matched, the target signal is The information is updated to the tracking target chain and the Kalman filter is corrected to determine whether the intersection ratio of the tracking target and the logistics line range is less than 0.2. If not, return to the initial detection process. If so, determine whether the target tracking is to the 4th frame and whether the initial position 4 frames ago is closer to the center area of ​​the logistics line (metal rod) than now. If so, determine whether there is a target in the tracking chain with a logistics line range intersection ratio greater than 0.1. If so, generate an alarm message and determine whether the time from the last alarm is more than 5s. If so, draw the latest n tracking target trajectories, save the abnormal transfer (NG) screen, and push the abnormal message to remind. When performing target detection, it can also detect whether there are more than N frames of acrylic board within M minutes. If so, save the NG screen and push the NG message.

[0176] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the battery transport detection method of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0177] This application also provides a battery transport detection device, please refer to Figure 8 , the battery transport detection device includes: The acquisition module 10 is used to acquire image data of multiple locations on the logistics line.

[0178] The detection module 20 is used to perform target detection on the image data to obtain the object position and object attributes.

[0179] The detection module 20 is also used to detect abnormal transportation of objects on the logistics line according to the location and properties of the objects to obtain detection results.

[0180] The battery transport detection device provided by the present application adopts the battery transport detection method in the above embodiment, which can solve the technical problem of low efficiency in detecting abnormal objects on the logistics line. Compared with the prior art, the beneficial effects of the battery transport detection device provided by the present application are the same as the beneficial effects of the battery transport detection method provided by the above embodiment, and the other technical features in the battery transport detection device are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0181] In one embodiment, the detection module 20 is further used to obtain the position of the rectangular frame of the object according to the position of the object; detect abnormal transportation of objects on the logistics line according to the position of the rectangular frame of the object and the object attributes to obtain the detection result.

[0182] In one embodiment, the detection module 20 is also used to determine a target to be tracked whose object attribute is a first tag; determine the first object rectangular frame position of the target to be tracked based on the object position, and determine the first rectangular frame according to the first object rectangular frame position; calculate the first overlap ratio between the first rectangular frame and the preset area of ​​the logistics line; when the first overlap ratio is less than the first preset ratio threshold, add the first object rectangular frame position to the list to be tracked; perform battery transfer detection on the target to be tracked based on the list to be tracked to obtain a detection result.

[0183] In one embodiment, the detection module 20 is also used to obtain the first rectangular frame position of the target to be tracked in the to-be-tracked list, and obtain the first rectangular frame according to the first rectangular frame position; determine the state of the target to be tracked according to the first rectangular frame; when the state of the target to be tracked is a preset state, determine that the target to be tracked has abnormal transportation, and obtain the detection result of abnormal battery transportation.

[0184] In one embodiment, the detection module 20 is also used to change the object attribute of the target to be tracked to a second label when the first overlap ratio between the first rectangular frame of the target to be tracked and the preset area of ​​the logistics line is greater than the second preset ratio threshold and the object attribute of the target to be tracked is the first label; obtain the initial position and the current position of the target to be tracked according to the list to be tracked; calculate the first distance between the initial position and the center position of the logistics line and the second distance between the current position and the center position of the logistics line; determine the number of image frames of the target to be tracked in the list to be tracked according to the first rectangular frame; determine the state of the target to be tracked according to the second label, the first distance, the second distance, the number of image frames and the first overlap ratio.

[0185] In one embodiment, the detection module 20 is further used to determine that the state of the target to be tracked is the preset state when the first distance is less than the second distance, the object attribute of the target to be tracked is changed to the second label, the first overlap ratio is less than the third preset ratio threshold, and the number of image frames reaches the set number of frames, the second preset ratio threshold is less than the third preset ratio threshold, and the first preset ratio threshold is greater than the third preset ratio threshold; and determine that the state of the target to be tracked is not the preset state when at least one of the following conditions is not met: the first distance is less than the second distance, the object attribute of the target to be tracked is changed to the second label, the first overlap ratio is less than the third preset ratio threshold, and the number of image frames reaches the set number of frames.

[0186] In one embodiment, the battery transport detection device further includes an update module; An updating module is used to predict the target to be tracked in the tracking list using a preset prediction strategy to obtain the predicted object rectangular frame position; obtain the predicted rectangular frame according to the predicted object rectangular frame position; calculate the overlap value between the predicted rectangular frame and the corresponding first rectangular frame in the tracking list; when the overlap value is greater than a second preset ratio threshold, determine that the predicted rectangular frame matches the first rectangular frame; when the predicted rectangular frame matches the first rectangular frame, update the first object rectangular frame position of the target to be tracked according to the predicted object rectangular frame position to obtain an updated tracking list.

[0187] In one embodiment, the update module is also used to determine that the predicted rectangular box does not match the first rectangular box when the overlap value is less than or equal to a second preset ratio threshold; when the predicted rectangular box does not match the first rectangular box, obtain the number of unmatched frames; when the number of unmatched frames is less than or equal to the preset frame number threshold, add the target corresponding to the predicted rectangular box to the to-be-tracked list for detection; when the number of unmatched frames is greater than the preset frame number threshold, remove the to-be-tracked target corresponding to the predicted rectangular box from the to-be-tracked list.

[0188] In one embodiment, the battery transport detection device further includes a reminder module; The reminder module is also used to provide a reminder of abnormal battery transportation when the detection result shows that the batteries on the logistics line are transported abnormally.

[0189] In one embodiment, the reminder module is also used to generate an alarm message and obtain the historical alarm interval time when the detection result shows that there is abnormal transportation of objects on the logistics line; when the historical alarm interval time exceeds the preset interval time, obtain the center point data of the target to be tracked; generate the trajectory information of the target to be tracked based on the center point data; send the alarm message and trajectory information to the user to remind the user of abnormal battery transportation.

[0190] In one embodiment, the reminder module is also used to determine a preset object whose object attribute is a third label; obtain a second rectangular box position of the preset object according to the object position, and determine the second rectangular box according to the second rectangular box position; calculate a second overlap ratio between the second rectangular box and the preset area of ​​the logistics line; determine that there is a preset object on the logistics line when the second overlap ratio is greater than a fourth preset ratio threshold; obtain the number of detections for detecting the presence of the preset object on the logistics line; when the preset object is detected within a preset time threshold and the number of detections is greater than a preset number threshold, determine that there is abnormal transportation of batteries on the logistics line, and obtain a detection result of abnormal battery transportation.

[0191] In one embodiment, the reminder module is also used to mark the preset object or the target to be tracked to obtain a marked image when the detection result shows that the battery on the logistics line has been abnormally transported; generate alarm information and obtain the historical alarm interval time; when the historical alarm interval time exceeds the preset interval time, send the alarm information and the marked image to the user to remind the user of abnormal battery transportation.

[0192] In one embodiment, the detection module 20 is further used to input the object position and object attributes into the anomaly detection model; and perform abnormal battery transportation detection on the logistics line through the anomaly detection model to obtain a detection result.

[0193] In one embodiment, the detection module 20 is further used to obtain training samples, wherein the training samples include normal battery transportation images and abnormal battery transportation images; the initial model is trained according to the training samples to obtain an abnormality detection model.

[0194] In one embodiment, the detection module 20 is also used to obtain original image data; annotate the original image data to obtain the original object position and original object attributes on the logistics line; track the objects on the logistics line according to the original object position and original object attributes to obtain normal battery transportation images and abnormal battery transportation images.

[0195] The present application provides a battery transport detection device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the battery transport detection method in the above-mentioned embodiment one.

[0196] Reference below Fig. 9 , which shows a schematic diagram of the structure of a battery transport detection device suitable for implementing the embodiment of the present application. The battery transport detection device in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Fig. 9 The battery transport detection device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0197] like Fig. 9As shown, the battery transport detection device may include a processing device 1001 (such as a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the battery transport detection device are also stored. The processing device 1001, ROM1002 and RAM1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the battery transport detection device to communicate with other devices wirelessly or wired to exchange data. Although the figure shows a battery transport detection device with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or have alternatively.

[0198] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0199] The battery transport detection device provided by the present application adopts the battery transport detection method in the above embodiment, which can solve the technical problem of low efficiency in detecting abnormal objects on the logistics line. Compared with the prior art, the beneficial effects of the battery transport detection device provided by the present application are the same as the beneficial effects of the battery transport detection method provided by the above embodiment, and the other technical features of the battery transport detection device are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0200] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0201] The above are only specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0202] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, and the computer-readable program instructions are used to execute the battery transport detection method in the above-mentioned embodiment.

[0203] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM: Random Access Memory), a read-only memory (ROM: Read Only Memory), an erasable programmable read-only memory (EPROM: Erasable Programmable Read Only Memory or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM: CD-Read Only Memory), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency: Radio Frequency), etc., or any suitable combination of the above.

[0204] The computer-readable storage medium may be included in the battery transport detection device; or it may exist independently without being assembled into the battery transport detection device.

[0205] The above-mentioned computer-readable storage medium carries one or more programs. When the above-mentioned one or more programs are executed by the battery transfer detection equipment, the battery transfer detection equipment is enabled to: obtain image data of multiple positions on the logistics line; perform target detection on the image data to obtain the object position and object properties; detect abnormal transfer of objects on the logistics line based on the object position and object properties to obtain detection results; and issue an abnormal transfer reminder when the detection result shows that there is abnormal transfer of objects on the logistics line.

[0206] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0207] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0208] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0209] The readable storage medium provided by the present application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned battery transport detection method, and can solve the technical problem of low efficiency in detecting abnormal objects on the logistics line. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the present application are the same as the beneficial effects of the battery transport detection method provided by the above-mentioned embodiment, and will not be elaborated here.

[0210] The present application also provides a computer program product, including a computer program, which implements the steps of the above-mentioned battery transport detection method when executed by a processor.

[0211] The computer program product provided by this application can solve the technical problem of low efficiency in detecting abnormal objects on logistics lines. Compared with the prior art, the beneficial effects of the computer program product provided by this application are the same as those of the battery transport detection method provided by the above embodiment, which will not be elaborated here.

[0212] The above are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect application in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A battery transport detection method, characterized in that: The battery transport detection method comprises: Acquire image data at multiple locations along the logistics line; Performing target detection on the image data to obtain the position and attributes of objects on the logistics line; The battery transportation on the logistics line is detected according to the object position and the object attribute to obtain a detection result, wherein the battery transportation on the logistics line includes: tracking the object according to the object position and the object attribute to determine whether there is abnormal battery transportation; When the detection result indicates that there is abnormal transportation of batteries on the logistics line, a reminder of abnormal transportation of batteries is issued.

2. The method according to claim 1, characterized in that The detecting of the battery transport on the logistics line according to the object position and the object attribute, and obtaining the detection result includes: Determine that the object attribute is a target to be tracked with a first tag; Determine a first object rectangular frame position of the target to be tracked based on the object position, and determine a first rectangular frame according to the first object rectangular frame position; Calculating a first overlapping ratio between the first rectangular frame and the preset area of ​​the logistics line; When the first overlap ratio is less than a first preset ratio threshold, adding the first object rectangular frame position to a to-be-tracked list; A battery transfer detection is performed on the target to be tracked based on the list to be tracked to obtain a detection result.

3. The method according to claim 2, characterized in that The performing battery transfer detection on the target to be tracked based on the list to be tracked to obtain the detection result includes: Acquire a first rectangular frame position of a target to be tracked in the to-be-tracked list, and obtain a first rectangular frame according to the first rectangular frame position; Determine the state of the target to be tracked according to the first rectangular frame; When the state of the target to be tracked is a preset state, it is determined that the target to be tracked has abnormal transport, and a detection result of abnormal transport of the battery is obtained.

4. The method according to claim 3, characterized in that Determining the state of the target to be tracked according to the first rectangular frame includes: When a first overlap ratio between the first rectangular frame of the target to be tracked and the preset area of ​​the logistics line is greater than a second preset ratio threshold and the object attribute of the target to be tracked is a first label, changing the object attribute of the target to be tracked to a second label; Obtaining the initial position and current position of the target to be tracked according to the list to be tracked; Calculate a first distance between the initial position and the center position of the logistics line and a second distance between the current position and the center position of the logistics line; Determine the number of image frames of the target to be tracked in the to-be-tracked list according to the first rectangular frame; The state of the target to be tracked is determined according to the second tag, the first distance, the second distance, the number of image frames, and the first overlapping ratio.

5. The method according to claim 4, characterized in that Determining the state of the target to be tracked according to the second tag, the first distance, the second distance, the number of image frames, and the first overlap ratio includes: Under the conditions that the first distance is less than the second distance, the object attribute of the target to be tracked is changed to the second label, the first overlap ratio is less than a third preset ratio threshold, and the number of image frames reaches a set number of frames, determining that the state of the target to be tracked is a preset state, the second preset ratio threshold is less than the third preset ratio threshold, and the first preset ratio threshold is greater than the third preset ratio threshold; When at least one of the following conditions is not satisfied: the first distance is less than the second distance, the object attribute of the target to be tracked is changed to the second label, the first overlap ratio is less than a third preset ratio threshold, and the number of image frames reaches a set number of frames, it is determined that the state of the target to be tracked is not a preset state.

6. The method according to claim 2, characterized in that The method further comprises: Predicting the target to be tracked in the to-be-tracked list to obtain the predicted rectangular frame position of the object; Obtaining a predicted rectangular frame according to the predicted object rectangular frame position; Calculate the overlap value between the predicted rectangular frame and the corresponding first rectangular frame in the to-be-tracked list; When the overlap value is greater than a second preset ratio threshold, determining that the predicted rectangular box matches the first rectangular box; In the case where the predicted rectangular frame matches the first rectangular frame, the position of the first object rectangular frame of the target to be tracked is updated according to the predicted object rectangular frame position to obtain an updated list to be tracked.

7. The method according to claim 6, characterized in that The method further comprises: When the overlap value is less than or equal to the second preset ratio threshold, determining that the predicted rectangular box does not match the first rectangular box; When the predicted rectangular frame does not match the first rectangular frame, obtaining the number of unmatched frames; When the number of unmatched frames is less than or equal to a preset frame number threshold, adding the target corresponding to the predicted rectangular box to a to-be-tracked list for detection; When the number of unmatched frames is greater than the preset frame number threshold, the target to be tracked corresponding to the predicted rectangular box is removed from the list to be tracked.

8. The method according to claim 1, characterized in that When the detection result indicates that the battery on the logistics line is transported abnormally, the abnormal battery transport reminder includes: When the detection result shows that there is abnormal transportation of batteries on the logistics line, an alarm message is generated and a historical alarm interval is obtained; When the historical alarm interval exceeds the preset interval, obtaining the center point data of the target to be tracked; Generating trajectory information of the target to be tracked according to the center point data; The alarm information and the trajectory information are sent to the user to remind the user of abnormal battery transportation.

9. The method according to claim 1, characterized in that The detecting of the battery transport on the logistics line according to the object position and the object attribute, and obtaining the detection result includes: Determining that the object attribute is a preset object of a third tag; Obtaining a second rectangular frame position of the preset object according to the object position, and determining a second rectangular frame according to the second rectangular frame position; Calculating a second overlapping ratio between the second rectangular frame and the preset area of ​​the logistics line; When the second overlapping ratio is greater than a fourth preset ratio threshold, determining that the preset object exists on the logistics line; Obtaining the number of detections of the presence of the preset object on the detection logistics line; When the preset object is detected within the preset time threshold and the number of detections is greater than the preset number threshold, it is determined that there is abnormal transportation of batteries on the logistics line, and a detection result of abnormal transportation of batteries is obtained.

10. The method according to any one of claims 1 to 9, characterized in that The method further comprises: When the detection result is that there is abnormal transportation of batteries on the logistics line, the preset object or the target to be tracked is marked to obtain a marked image; Generate alarm information and obtain historical alarm intervals; When the historical alarm interval exceeds the preset interval, the alarm information and the annotated image are sent to the user to remind the user of abnormal battery transportation.

11. The method according to claim 1, characterized in that The detecting of the battery transport on the logistics line according to the object position and the object attribute, and obtaining the detection result includes: inputting the object position and the object attributes into an anomaly detection model; The abnormal transportation detection of batteries on the logistics line is performed using the abnormal detection model to obtain a detection result.

12. The method according to claim 11, characterized in that The anomaly detection model is trained by: Acquire training samples, wherein the training samples include normal battery transport images and abnormal battery transport images; The initial model is trained according to the training samples to obtain an anomaly detection model.

13. The method according to claim 12, characterized in that The obtaining of training samples comprises: Get the original image data; Annotate the original image data to obtain the original object position and original object attributes on the logistics line; The objects on the logistics line are tracked according to the original object positions and the original object attributes to obtain normal battery transportation images and abnormal battery transportation images.

14. A battery transport detection device, characterized in that: The device comprises: An acquisition module, used to acquire image data of multiple locations on the logistics line; A detection module, used to perform target detection on the image data to obtain the position and properties of objects on the logistics line; The detection module is further used to detect the battery transportation on the logistics line according to the object position and the object attributes to obtain a detection result, wherein the battery transportation on the logistics line includes: tracking the object according to the object position and the object attributes to determine whether there is abnormal battery transportation; The reminder module is used to provide a reminder of abnormal battery transportation when the detection result shows that the batteries on the logistics line are transported abnormally.

15. A battery transport detection device, characterized in that: The device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the battery transport detection method according to any one of claims 1 to 13.

16. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the battery transport detection method according to any one of claims 1 to 13 are implemented.

Images