Iron and steel logistics event management method and device, electronic equipment and storage medium

The event management method in steel logistics uses a pre-trained model to automate object detection and tracking, improving accuracy and efficiency by generating comprehensive feature information and event management results, addressing inefficiencies and misjudgment in large-scale scenarios.

CN119067542APending Publication Date: 2024-12-03欧冶云商股份有限公司
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411037648.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-30
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

Existing methods for event management in steel logistics, particularly in large-scale and complex scenarios, suffer from low efficiency and reliability due to manual video browsing and inadequate network conditions, leading to inefficiencies and increased risks of misjudgment.

Method used

An event management method utilizing a pre-trained target detection model to identify and track objects in multiple frames of steel logistics images, generating comprehensive feature information and event management results through image analysis and advanced tracking algorithms, thereby automating the process.

Benefits of technology

This approach enhances the accuracy and efficiency of target detection, reduces manual intervention, and enables real-time, intelligent event management, quickly responding to logistics events and minimizing misjudgment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119067542A_ABST
    Figure CN119067542A_ABST
Patent Text Reader

Abstract

The invention provides an event management method and device for steel logistics, electronic equipment and a storage medium, and relates to the technical field of steel enterprise informatization. The event management method for steel logistics comprises the steps of determining object detection data corresponding to a target detection object in a steel logistics image according to multiple frames of steel logistics images and a pre-trained target detection model; performing image analysis on the corresponding multiple frames of steel logistics images based on the object detection data to determine target feature information corresponding to the target detection object; generating a target object tracking result based on the target feature information, the object detection data and the multiple frames of steel logistics images; and generating event management result information according to the event management requirement and the target object tracking result. According to the invention, continuous tracking of the target object in the logistics process can be realized, various events occurring in the logistics process can be quickly responded, the event processing time is shortened, the influence of the events on the logistics process is reduced, and the event management efficiency is integrally improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of information technology for steel enterprises, for example, to an event management method, device, electronic equipment and storage medium for steel logistics. Background Art

[0002] Steel logistics is a cross-industry that uses "steel" as a carrier, "logistics" as an operating method, and "information" as the core. The steel logistics scene mainly includes the transportation, warehousing, loading and unloading, handling, packaging, and processing of steel. Due to the large scale, frequent operations, and complex operation processes of the steel logistics scene, the supervision requirements for storage areas in the steel logistics scene are relatively high. With the advancement of social science and technology, the steel logistics scene has gradually introduced advanced technologies such as big data, cloud computing, and the Internet of Things to manage events in storage areas.

[0003] At present, a large number of monitoring devices are installed in the storage area of ​​steel logistics, and the staff can manage all aspects of steel logistics through monitoring devices. However, the method of manually browsing videos cannot meet the requirements of safety and reliability of the entire process of steel logistics. Therefore, in related technologies, many manufacturing companies have begun to introduce machine vision technology and manage events through cloud deployment. However, the cloud deployment model has low event management efficiency in steel logistics scenarios with large scale, frequent operations, poor network conditions, etc. Summary of the invention

[0004] The present application aims to provide an event management method, device, electronic device and storage medium for steel logistics.

[0005] According to one aspect of the present application, a method for event management of steel logistics is proposed, comprising: determining object detection data corresponding to a target detection object in a steel logistics image based on multiple frames of steel logistics images and a pre-trained target detection model; performing image analysis on the corresponding multiple frames of steel logistics images based on the object detection data to determine target feature information of the corresponding target detection object; generating a target object tracking result based on the target feature information, the object detection data and the multiple frames of steel logistics images; and generating event management result information according to event management requirements and the target object tracking result.

[0006] According to one aspect of the present application, an event management device for steel logistics is proposed, comprising: a data determination module, used to determine object detection data corresponding to a target detection object in a steel logistics image based on multiple frames of steel logistics images and a pre-trained target detection model; a feature determination module, used to perform image analysis on the corresponding multiple frames of steel logistics images based on the object detection data to determine target feature information of the corresponding target detection object; a tracking result determination module, used to generate a target object tracking result based on the target feature information, the object detection data and the multiple frames of steel logistics images; and an event information storage module, used to generate event management result information according to event management requirements and the target object tracking result.

[0007] According to one aspect of the present application, an electronic device is provided, which includes: a processor; and a memory storing a computer program, wherein when the computer program is executed by the processor, the processor executes the event management method for steel logistics as described above.

[0008] According to one aspect of the present application, a non-transitory computer-readable medium is provided, on which readable instructions are stored. When the instructions are executed by a processor, the processor executes the event management method for steel logistics as described above.

[0009] It should be understood that the foregoing general description and the following detailed description are exemplary only and are not restrictive of the present application.

[0010] Beneficial effects:

[0011] Through the above-mentioned embodiments provided by the present application, the target detection object can be quickly and accurately identified from multiple frames of steel logistics images using a pre-trained target detection model, significantly improving the accuracy and efficiency of target detection. This helps to reduce manual intervention and reduce the risk of misjudgment and missed judgment. Based on the in-depth analysis of multiple frames of images based on object detection data, the feature information of the target detection object can be more comprehensively extracted, and these feature information can provide rich data support for subsequent target object tracking and event management. By combining target feature information, object detection data and multiple frames of images, and using advanced tracking algorithms, continuous tracking of target objects in the logistics process can be achieved. According to event management requirements and target object tracking results, event management result information is automatically generated, which realizes the intelligence and automation of event management, and can quickly respond to various events occurring in the logistics process, reduce event processing time, and reduce the impact of events on the logistics process. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without exceeding the scope of protection required by the present application.

[0013] Figure 1 A schematic diagram of an application scenario provided for an embodiment of the present application;

[0014] Figure 2 A flowchart of an event management method for steel logistics provided in an embodiment of the present application;

[0015] Figure 3 A flowchart of pre-processing the steel logistics monitoring data in step S20 provided in an embodiment of the present application to determine multiple frames of steel logistics images;

[0016] Figure 4 A flowchart for implementing step S20 provided in an embodiment of the present application;

[0017] Figure 5 A flowchart for implementing step S21 provided in an embodiment of the present application;

[0018] Figure 6 A flowchart for implementing step S22 provided in an embodiment of the present application;

[0019] Figure 7 Another implementation flow chart of step S22 provided in an embodiment of the present application;

[0020] Figure 8 A flowchart for implementing step S23 provided in an embodiment of the present application;

[0021] Fig. 9 A flowchart of another event management method for steel logistics provided in an embodiment of the present application;

[0022] Fig.10 A flowchart of another event management method for steel logistics provided in an embodiment of the present application;

[0023] Fig.11 A block diagram of an event management device for steel logistics provided in an embodiment of the present application;

[0024] Fig.12 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0025] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0026] Figure 1 Schematic diagram of application scenario of the embodiment of the present application. Steel logistics sites may include specific locations such as warehouses and steel transportation. Figure 1 In the scenario shown, the event management method for steel logistics can be built on the event management edge server. Multiple monitoring devices can be set up in the steel logistics site to monitor various operation processes in the steel logistics site, generate steel logistics monitoring data and send it to the event management edge server, so that the event management edge server can manage various events corresponding to the aforementioned operation processes.

[0027] For specific implementation methods, please refer to the following embodiments.

[0028] Figure 2 This is a flow chart of the event management method for steel logistics provided in the embodiment of the present application. The method of this embodiment can be applied to the event management edge server. Figure 2 As shown, the method includes: step S20, step S21, step S22 and step S23.

[0029] Step S20, determining object detection data corresponding to the target detection object in the steel logistics image according to the multiple frames of steel logistics image and the pre-trained target detection model.

[0030] In the present application, multiple frames of steel logistics images may be taken by monitoring equipment installed in a steel logistics site, and multiple consecutive frames of steel logistics images may constitute a piece of video data. The pre-trained target detection model may be a model that is pre-trained and stored based on the architecture of a public detection model and in accordance with actual needs for use in a steel logistics scenario. The model has the ability to input image data and directly output object detection data.

[0031] In some implementations, a model training platform can be pre-set to store data sets required for various model training and various public model frameworks. The target detection model required for this application can be trained on the model training platform and sent to the event management edge server through the Internet of Things platform.

[0032] The target detection object in the steel logistics image can be used to represent the object to be tracked in a certain frame of the steel logistics image, or the object involved in event processing. The target detection object can include people, steel coils, steel materials, vehicles, etc.

[0033] The object detection data may be the location information of the target detection object in the image and the name label of the object determined after image analysis, wherein the name label of the object may include a label of a person, a label of a steel coil, a label of a steel material, a label of a vehicle, etc.

[0034] According to the example embodiment, a pre-trained target detection model issued by the Internet of Things platform can be obtained first, and a detection model with the best performance and the latest version can be selected, such as the Yolov8 target detection model. Then, the pre-trained target detection model can be used to identify target detection objects such as people and steel coils in multiple frames of steel logistics images and corresponding object detection data.

[0035] Step S21 , performing image analysis on the corresponding multiple frames of steel logistics images based on the object detection data to determine target feature information of the corresponding target detection object.

[0036] In the present application, target feature information can be used to characterize information on unique attributes of the target detection object, such as appearance features, which can include shape, color, size, etc. In some implementations, it can also include movement features.

[0037] According to the example embodiment, the position of the target detection object in the steel logistics image can be extracted from the object detection data. By performing image analysis on the corresponding positions of multiple frames of steel logistics images, the image features of the target detection object can be roughly obtained as target feature information.

[0038] Step S22, generating a target object tracking result based on the target feature information, the object detection data and the multiple frames of steel logistics images.

[0039] In the present application, the target object tracking results can be used to characterize the continuous position, motion trajectory and other information of the target detection object over a period of time obtained by comprehensive analysis and processing of target feature information, object detection data and multiple frames of steel logistics images.

[0040] According to the example embodiment, for each frame of the multiple frames of steel logistics images, the position of the specific target detection object contained in the image can be determined based on the target feature information, and then the multiple frames of steel logistics images corresponding to the same target detection object can be extracted. Then, by using the object detection data, the position change of the same target detection object in the corresponding multiple frames of steel logistics images can be determined to obtain the target object tracking result.

[0041] Step S23: Generate event management result information according to the event management requirements and the target object tracking results, and store the event management result information.

[0042] In this application, event management requirements can represent the special requirements of managers in steel logistics scenarios when tracking targets. For example, for a certain target detection object, you want to obtain some special data. If the target detection object is a vehicle, while tracking the vehicle's trajectory, you may need to obtain data such as the license plate number.

[0043] The event management result information may be overall data obtained by combining event management requirements and target object tracking results, and may include the movement trajectory of the target detection object, multiple data corresponding to the requirements, and the like.

[0044] According to the example embodiment, the event management requirements may be analyzed to determine which data is needed. Then, the data may be supplemented and acquired in accordance with the aforementioned image analysis method. Then, the supplemented data and the target object tracking result are determined as event management result information as a whole and stored in a preset server.

[0045] This application uses a pre-trained target detection model to quickly and accurately identify target detection objects from multiple frames of steel logistics images, significantly improving the accuracy and efficiency of target detection. This helps to reduce manual intervention and reduce the risk of misjudgment and missed judgment. Based on the in-depth analysis of multiple frames of images based on object detection data, it is possible to more comprehensively extract the feature information of the target detection object, which can provide rich data support for subsequent target object tracking and event management. By combining target feature information, object detection data and multiple frames of images, and using advanced tracking algorithms, it is possible to achieve continuous tracking of target objects in the logistics process. According to event management requirements and target object tracking results, event management result information is automatically generated, realizing the intelligence and automation of event management, and can quickly respond to various events occurring in the logistics process, reduce event processing time, and reduce the impact of events on the logistics process.

[0046] According to some embodiments, before step S20, multiple frames of steel logistics images can also be determined by receiving steel logistics monitoring data sent by monitoring equipment, specifically: receiving steel logistics monitoring data; wherein the steel logistics monitoring data is detected and sent by monitoring equipment installed in a steel logistics site; and pre-processing the steel logistics monitoring data to obtain multiple frames of steel logistics images.

[0047] According to an example embodiment, multiple monitoring devices, such as high-definition cameras, may be installed in a steel logistics site. The types, versions, and performances of these monitoring devices may not be exactly the same. The monitoring device may send the captured steel logistics monitoring data to an event management edge server (hereinafter referred to as a server), which may include relevant data of the operation process in the steel logistics site and may also include a device ID representing the monitoring device. The server may perform pre-processing operations such as data cleaning and standardization on the steel logistics monitoring data to obtain multiple frames of steel logistics images.

[0048] This application realizes real-time monitoring of the logistics process by installing a variety of monitoring equipment in the steel logistics site, ensuring the integrity and real-time nature of the data. The monitoring equipment covers key logistics nodes, effectively reduces data blind spots, and improves the comprehensiveness and accuracy of detection data. Pre-processing of steel logistics monitoring data can provide a high-quality data source for subsequent data analysis and processing.

[0049] According to some embodiments, reference Figure 3 In step S20, the steel logistics monitoring data is pre-processed to determine multiple frames of steel logistics images, which can be specifically achieved through steps S300, S301 and S302.

[0050] In step S300, a plurality of first images to be processed are extracted from the steel logistics monitoring data.

[0051] In the present application, the first image to be processed may represent an image that is initially screened out from the steel logistics monitoring data and needs to be further processed or analyzed.

[0052] According to an example embodiment, the steel logistics monitoring data may be split into a plurality of adjacent first images to be processed according to the time when the data was generated.

[0053] According to some embodiments, the first image to be processed may be selected starting from the first scene of the steel logistics monitoring data according to a preset screening condition, such as a mode of taking one image every few seconds.

[0054] According to other embodiments, the preset screening condition may also be to first determine the corresponding monitoring equipment in a certain logistics link or a certain event type. Based on the unique equipment ID of the monitoring equipment, the detection data corresponding to these monitoring equipment are first selected from the steel logistics monitoring data, and then the data is divided into frames of first images to be processed.

[0055] In step S301, a plurality of first images to be processed are decoded according to the equipment performance of the monitoring equipment corresponding to the steel logistics monitoring data, and the decoded plurality of first images to be processed are converted into second images to be processed in a required format.

[0056] In this application, equipment performance may refer to the technical specifications and capabilities of monitoring equipment installed in a steel logistics site, including but not limited to resolution, frame rate, compression algorithm, etc. These performance parameters may directly affect the quality of captured images and the complexity of subsequent processing. In some implementations, a comprehensive performance analysis of the monitoring equipment used in a steel logistics site may be performed in advance to determine key parameters such as resolution, frame rate, compression algorithm, etc. of each device. Based on the analysis results, an equipment performance database is established so that it can be quickly queried and referenced when processing images.

[0057] The required format may be a specific format requirement set for the image data according to the requirements of subsequent processing or application. This required format may be determined by image analysis software, database or display platform, etc. In some implementations, a picture size format with a resolution of 640*640 may be selected. In addition, color and other requirements may also be included.

[0058] The second image to be processed may be used to represent an image that, after being decoded and format converted, meets specific required format requirements and is ready for further processing or analysis.

[0059] According to the example embodiment, the pre-stored data related to the device performance of the monitoring device can be obtained to obtain the resolution, frame rate and compression algorithm of the monitoring device. For each first image to be processed, according to the performance parameters of the monitoring device from which it comes, a suitable decoding algorithm is selected for decoding. Then, the decoded first image to be processed can be converted into a required format using an image processing library or tool.

[0060] In step S302, an image enhancement operation is performed on the second image to be processed to obtain multiple frames of steel logistics images.

[0061] In this application, image enhancement can be a process of improving image quality through a series of image processing techniques to make it more suitable for subsequent analysis or visual perception. Image enhancement techniques include but are not limited to brightness adjustment, contrast enhancement, noise suppression, sharpening, etc.

[0062] According to the example embodiment, each second image to be processed can be subjected to a preliminary quality assessment to identify possible quality problems, such as insufficient brightness, low contrast, excessive noise, etc. According to the assessment results, a suitable image enhancement technology is selected for preprocessing. For example, for images with insufficient brightness, a brightness adjustment technology can be used; for images with low contrast, a contrast enhancement technology can be used; for images with excessive noise, a noise suppression technology can be used. Then, the second image to be processed after image enhancement is determined as a multi-frame steel logistics image.

[0063] This application significantly improves the processing efficiency of steel logistics monitoring data through data extraction, decoding, format conversion and image enhancement processes. Through image enhancement operations and continuity checks, the high quality and continuity of multiple frames of steel logistics images are ensured, providing a reliable data basis for subsequent analysis. This pre-processing process can flexibly respond to steel logistics monitoring data from different sources and requirements, improving the adaptability and scalability of the method of this application.

[0064] According to some embodiments, reference Figure 4 , step S20 can be specifically implemented through step S200 and step S201.

[0065] In step S200, a pre-trained optional detection model is searched to determine the latest version of the optional detection model as the target detection model.

[0066] In this application, the optional detection model can be a series of pre-trained models that can be used for target detection and are trained and updated in the above-mentioned model training platform. These models may have different architectures, performances, and characteristics and can be used for different target detection requirements.

[0067] According to an example embodiment, the model training platform may mark the corresponding update time when updating the optional detection model, and may select the model with the shortest time interval from the current moment from the pre-trained optional detection models as the target detection model.

[0068] In step S201, multiple frames of steel logistics images are input into a target detection model to determine object detection data; wherein the object detection data includes the position coordinates and name labels of the target detection objects in the multiple frames of steel logistics images.

[0069] In this application, object detection data can represent detailed information about the target detection object (such as steel, transport vehicles, etc.) obtained after the target detection model processes multiple frames of steel logistics images, including their position coordinates and name labels in each frame of the image. Among them, the position coordinates can be used to describe the numerical information of the specific position of the target detection object in the image, usually including the horizontal coordinate and the vertical coordinate. The name label can be a text label associated with the target detection object, which is used to describe the category or name of the object.

[0070] According to an example embodiment, multiple frames of steel logistics images are input into a target detection model, and object detection data is directly output.

[0071] This application uses the latest version of the pre-trained optional detection model as the target detection model, ensuring that the model can utilize the latest research results and technical optimizations, thereby improving the accuracy of target detection object recognition. It can accurately identify target objects in complex steel logistics scenarios and reduce false positives and false negatives. Continuously inputting multiple frames of steel logistics images into the target detection model can generate object detection data in real time.

[0072] According to some embodiments, reference Figure 5 , step S21 can be specifically implemented through step S210 and step S211.

[0073] In step S210, multiple frames of local images corresponding to the position coordinates are selected from the corresponding multiple frames of steel logistics images.

[0074] In the present application, the local image may be a smaller image cropped from the original image and containing a specific area or object.

[0075] According to the example embodiment, since the position coordinates are the specific positions of the target detection object in the steel logistics image, each position of the steel logistics image itself corresponds to the coordinates. Based on the position coordinates, a suitable area can be selected on each frame of the steel logistics image, and the area needs to be able to completely contain the target detection object. The area is determined as the local image corresponding to the steel logistics image.

[0076] In step S211, multiple frames of local images are analyzed to determine target feature information of the target detection object; wherein the target feature information includes appearance features and motion features.

[0077] In this application, the appearance feature can be used to describe the information of the appearance characteristics of the target detection object, such as color, shape, texture, etc. The motion feature can be used to describe the information of the motion state and behavior pattern of the target detection object, such as speed, acceleration, etc.

[0078] According to an example embodiment, computer vision techniques such as feature point detection, edge detection, texture analysis, etc. can be used to extract the apparent features of the target detection object from each frame of the local image. These features may include color histograms, shape descriptors (such as contours, corners), texture features (such as local binary patterns LBP), etc. The temporal correlation between multiple frames of local images can then be used to analyze the motion characteristics of the target detection object. Use optical flow, feature point tracking, or background subtraction-based methods to track the position changes of the target detection object in consecutive frames. Calculate the motion characteristics of the target such as speed, acceleration, and trajectory based on the position change.

[0079] In some implementations, the extracted surface features are stored for subsequent feature matching or classification.

[0080] Based on the position coordinates in the object detection data, this application can accurately select local images containing the target detection object from multiple frames of steel logistics images. This accuracy ensures the accuracy and effectiveness of subsequent feature extraction and avoids interference from background noise and irrelevant information. By analyzing multiple frames of local images, not only can the apparent features of the target detection object be extracted, but also the motion features can be analyzed in combination with the time dimension. This comprehensive feature analysis provides rich data support for a deep understanding of the behavior patterns and state changes of the target detection object.

[0081] According to some embodiments, reference Figure 6 Step S22 can be specifically implemented through steps S220, S221, S222 and S223.

[0082] In step S220, a pre-trained target tracking model is obtained; target feature information is input into the target tracking model to determine a plurality of frames of optional tracking images containing the target detection object;

[0083] In this application, the pre-trained target tracking model may refer to a machine learning model that has been trained on the above-mentioned model training platform through a large amount of data and can automatically identify and track specific targets in a video or image sequence. The model can use the input target feature information to locate and track the target object in continuous image frames.

[0084] The multiple frames of optional tracking images can represent a series of image frames that contain target detection objects and meet the tracking requirements after being processed by the target tracking model. These image frames provide a basis for subsequent further analysis or processing.

[0085] According to an example embodiment, a pre-trained target tracking model issued by an Internet of Things platform can be received and obtained. Then, the target feature information is input into the target tracking model, and the model can use the feature information to search and track the target detection object in continuous image frames, and then output a series of image frame position information containing the target detection object. According to the image frame position information, the corresponding image frames can be extracted from the original multi-frame steel logistics image to form a multi-frame optional tracking image.

[0086] In step S221, the similarity of the target detection objects in the multiple frames of optional tracking images is calculated based on the target feature information and a preset multi-target tracking algorithm.

[0087] In this application, the multi-target tracking algorithm can process a video or image sequence containing multiple moving targets and continuously track the position and state of these targets in the sequence. The algorithm can combine target detection, data association and state estimation techniques.

[0088] Similarity can be used to quantify the degree of similarity between two or more target detection objects. In multi-target tracking, similarity calculation is crucial to distinguish different targets, avoid identity exchange, and maintain the continuity of target trajectories.

[0089] According to an example embodiment, for each target detection object in each frame of image, a preset multi-target tracking algorithm is used to calculate its similarity with all possible corresponding targets in the previous frame. The similarity calculation can be based on a variety of factors, such as the Euclidean distance, cosine similarity, Mahalanobis distance, etc. of the feature vector, depending on the type and nature of the target feature information.

[0090] According to some embodiments, a similarity matrix may be constructed, in which each element represents the similarity between two objects in adjacent frames.

[0091] In step S222, images corresponding to abnormal movement trajectories are eliminated from the multiple frames of optional tracking images according to the similarity, so as to determine the target movement speed of the multiple frames of target tracking images and the target detection object.

[0092] In this application, abnormal movement trajectories may refer to trajectories that are significantly inconsistent with the normal movement pattern of the target detection object. These trajectories may be caused by misidentification, occlusion, noise, recognition errors caused by objects being too similar, or other interference factors.

[0093] The target moving speed can be used to indicate the distance that the target detection object moves in a specific direction per unit time. In multi-frame images, the moving speed of the target can be determined by calculating the position change and time interval between consecutive frames.

[0094] According to the example embodiment, the similarity matrix can be analyzed to find those tracks with significantly low similarity to the surrounding targets, which may indicate abnormal movement or misidentification, and the corresponding images can be eliminated. Then, a data association algorithm, such as the Hungarian algorithm, the network flow algorithm, etc., can be used in combination with the similarity matrix to segment and reorganize the target track, and eliminate the images corresponding to the abnormal movement track, as the corresponding target detection objects in these images may not be the same.

[0095] After eliminating abnormal moving trajectories, the remaining image frames constitute a multi-frame target tracking image. In the multi-frame target tracking image, the positions of the target detection objects in consecutive frames are selected.

[0096] Calculate the pixel distance of the target position between adjacent frames and convert it to the actual distance based on the parameters of the monitoring device (such as focal length, resolution, etc.). Calculate the time interval between adjacent frames (usually in seconds).

[0097] The formula "speed = distance / time" can be used to calculate the moving speed of the target detection object. The above calculation can be performed on multiple consecutive frame pairs and the average value can be taken to smooth the speed changes and obtain a more stable target moving speed estimate.

[0098] In step S223, a unique code is set for the target detection object according to the multiple frames of target tracking images; and a target object tracking result is generated according to the unique code, the object detection data and the target moving speed.

[0099] In the present application, a unique identifier may be assigned to each target detection object. This code is used to uniquely identify and reference the target object during the entire tracking process, and the code may be used as a unique code for the target detection object. The target object tracking result may be a result report or data set that is generated based on the unique code, object detection data, and target movement speed, and describes the behavior and state of the target detection object during the entire tracking process.

[0100] According to an example embodiment, a unique code is assigned to each target detection object that appears continuously in multiple frames of target tracking images. This code can be generated based on a timestamp, a sequence number, or other uniqueness-guaranteed algorithm. In the database of the target tracking system, a corresponding record is created for each unique code to store all relevant information of the target detection object.

[0101] According to the unique code, all relevant information of the target detection object, including detection data and movement speed, is retrieved from the database. This information is sorted and generated to generate the target object tracking result. This result can be a detailed report containing key information such as the identity information, movement trajectory, speed change, and residence time of the target detection object. The tracking results can be presented in various forms, such as text reports, charts, animations, or real-time interface displays, depending on the actual application scenario and requirements.

[0102] This application uses a pre-trained target tracking model and combines it with the precise input of target feature information to effectively filter out frames containing target detection objects from a large number of steel logistics images, greatly reducing the possibility of misidentification and missed identification, thereby improving tracking accuracy. Using a preset multi-target tracking algorithm combined with similarity calculation, it can accurately identify and eliminate images corresponding to abnormal movement trajectories, ensuring that the trajectory of the target detection object is continuous and stable, thereby improving the accuracy of the target object tracking results.

[0103] According to some embodiments, reference Figure 7 Step S22 can also perform post-processing on multiple frames of steel logistics images, which can be specifically achieved through steps S224 and S225.

[0104] In step S224, data formatting is performed on the target object tracking result.

[0105] In this application, data formatting can mean converting and processing raw data according to certain rules and structures so that it has a unified format and type, which is convenient for subsequent storage, analysis and application. The data formatting standard can be pre-set. For the target object tracking results, the pre-set data fields include: target object ID (string), location coordinates (latitude and longitude, floating point type), speed (meters / second, floating point type), timestamp (record tracking time, timestamp format), etc.

[0106] According to the example embodiment, one or more formatted data structures are constructed to store the converted data according to the pre-set formatting standards. These data structures can be simple data structures (such as arrays, dictionaries) or complex data models (such as objects, classes), and then corresponding scripts or programs are written to perform format conversion.

[0107] In step S225, according to the formatted target object tracking result, multiple frames of steel logistics images are tracked and annotated to generate multiple frames of annotated event management images.

[0108] In the present application, tracking and labeling can be based on the target object tracking results, marking the location, name, etc. of the target monitoring object in multiple frames of images, and the completion of tracking and labeling can be specifically indicated by a label displayed on the image.

[0109] According to an example embodiment, a marking box (such as a rectangular box, a circular box, etc.) may be drawn around the located target object. Additional annotation information, such as a unique code, a timestamp, a speed, etc., may be added inside or next to the marking box. The above operations are performed one by one on multiple frames of images to ensure that the target object in each frame of the image is correctly annotated. The annotated image is managed as a multi-frame annotation event image.

[0110] This application ensures data consistency and standardization by formatting the target object tracking results. This facilitates subsequent data analysis, storage, and sharing, and reduces compatibility issues caused by inconsistent data formats. Based on the formatted tracking results, multiple frames of steel logistics images are tracked and annotated, and the generated multi-frame annotated event management images directly display the key positions and state changes of the target detection objects in the logistics process. This intuitive image display method helps to quickly identify abnormal situations and improve the speed of problem response.

[0111] According to some embodiments, reference Figure 8 , step S23 can be implemented through step S230 and step S231.

[0112] In step S230, an event management requirement is received and analyzed to obtain a corresponding new event result.

[0113] In this application, the newly added event result can be a new processing logic for a specific type of event determined after analysis based on event management requirements, so that a corresponding result record can be generated or a specific operation can be performed when an event occurs.

[0114] According to an example embodiment, an event management requirement sent by a manager may be received, and newly added points of interest may be analyzed, such as a vehicle license plate number, etc. Based on the newly added points of interest, a corresponding pre-trained model may be used to output corresponding newly added event results.

[0115] In step S231, the newly added event result, the target object tracking result and the multiple-frame annotated event management images are determined as event management result information.

[0116] According to an example embodiment, the event adding result, the target object tracking result and the multi-frame annotated event management image are integrated as event management result information and stored in a corresponding server or database.

[0117] By receiving and analyzing event management needs, this application enables the system to flexibly respond to special events or changes in demand in various logistics scenarios, ensuring the timeliness and pertinence of management decisions, and helping to better respond to management needs. The newly added event results are combined with the target object tracking results and multi-frame annotated event management images to generate event management result information. This process achieves rapid integration and presentation of information. This enables managers to quickly understand the full picture of the event and make decisions in a timely manner, improving problem response speed and processing efficiency. Storing event management result information not only facilitates subsequent data analysis and query, but also provides strong data support for the company's long-term development and strategic planning.

[0118] According to some embodiments, reference Fig. 9 The above-mentioned time management method for steel logistics may further include storing event management result information, which may be implemented through step S24 and step S25.

[0119] In step S24, the multi-frame annotated event management image is sent to the first database for storage through the HTTP protocol, and the storage path of the multi-frame annotated event management image is stored.

[0120] HTTP is the Hypertext Transfer Protocol, the most widely used network protocol on the Internet. It is a transmission protocol used to transfer hypertext from the World Wide Web (WWW) server to the local browser.

[0121] In this application, the storage path may refer to the specific location or address of data on a device such as a hard disk or a database, which is used for subsequent retrieval or access to the data.

[0122] According to an example embodiment, a first database is pre-established, and a preset programming language and HTTP client library are selected to construct a POST request. The multi-frame annotated event management image can be attached as a binary file and sent together with the POST request to an API endpoint on a server hosting the first database. The hosted server parses the request and the binary file, and then saves the image to the first database. The hosted server can then feed back the storage path of the image to the event processing edge server, and the event processing edge server stores it accordingly.

[0123] In step S25, the newly added event results and the target object tracking results are sent to the second database for storage through the MQTT protocol.

[0124] MQTT is a message queue telemetry transmission protocol. It is a lightweight message transmission protocol based on the publish / subscribe model. It is widely used in the field of Internet of Things for communication between devices.

[0125] According to an example embodiment, an MQTT topic may be defined for the newly added event result and the target object tracking result. A preset client library supporting the MQTT protocol, such as a Python client library, is used to encapsulate the newly added event result and the target object tracking result into an MQTT message. The MQTT message is sent to a client of the second database so that the client parses the message and inserts the newly added event result and the target object tracking result into the second database for storage, which may include creating or updating corresponding tables, records, etc.

[0126] This application uses the HTTP protocol to transmit multi-frame annotated event management images. The protocol has high stability and reliability in file transfer and can ensure the integrity and accuracy of image data. At the same time, the HTTP protocol supports large file transfer and is suitable for processing multi-frame image data. New event results and target object tracking results are sent through the MQTT protocol. The protocol is lightweight and low-overhead, and is particularly suitable for data transmission in scenarios such as the Internet of Things and mobile devices. The high efficiency of the MQTT protocol ensures the real-time and rapid response of these key information. Storing the multi-frame annotated event management images in the first database and saving their storage path helps to achieve classified management and rapid retrieval of data.

[0127] According to some embodiments, reference Fig.10 The above-mentioned time management method for steel logistics may also include generating an event record of complete event information, which may be implemented through steps S26 and S27.

[0128] In step S26, the storage paths of the multi-frame annotated event management images are merged with the newly added event results to generate complete event information.

[0129] In this application, merging may refer to integrating the storage paths and newly added event results together to form a more complete data set or report.

[0130] According to an example embodiment, the storage path of the multi-frame annotated event management image is integrated with the newly added event result to generate complete event information.

[0131] In step S27, the complete event information is stored in the second database, and a corresponding event record is generated.

[0132] According to an example embodiment, the complete event information is sent to a second database for storage, and an event record is generated. In some implementations, a unique identifier may be set for each event record.

[0133] This application generates complete event information including image data and event description information by merging the storage path of multi-frame annotated event management images with the newly added event results. This data integration method ensures the comprehensiveness and consistency of the data and avoids the risk of information fragmentation and omissions. The generation of complete event information establishes a clear association between image data and event results. This association helps managers to intuitively understand the specific details of the event when viewing event records, improving the efficiency of information understanding and decision-making. The complete event information is uniformly stored in the second database and the corresponding event records are generated, which simplifies the data management process. Managers can quickly retrieve the required event information by querying the database without switching between multiple data sources or performing complex data integration operations.

[0134] The following describes an apparatus embodiment of the present application, which can be used to execute the method embodiment of the present application. For details not disclosed in the apparatus embodiment of the present application, reference can be made to the method embodiment of the present application.

[0135] Fig.11 This is a block diagram of an event management device for steel logistics provided in an embodiment of the present application. Fig.11 As shown, the event management device 1100 for steel logistics includes a data determination module 1101 , a feature determination module 1102 , a tracking result determination module 1103 and an event information storage module 1104 .

[0136] The data determination module 1101 can determine the object detection data corresponding to the target detection object in the steel logistics image according to multiple frames of steel logistics images and a pre-trained target detection model.

[0137] The feature determination module 1102 may perform image analysis on the corresponding multiple frames of steel logistics images based on the object detection data to determine target feature information of the corresponding target detection object.

[0138] The tracking result determination module 1103 may generate a target object tracking result based on the target feature information, the object detection data and the multiple frames of steel logistics images.

[0139] The event information generation module 1104 may generate event management result information according to event management requirements and target object tracking results.

[0140] Optionally, the steel logistics event management device 1100 further includes a steel logistics image acquisition module 1105. The steel logistics image acquisition module 1105 can receive steel logistics monitoring data; the steel logistics monitoring data here is detected and sent by monitoring equipment installed in the steel logistics site; and pre-process the steel logistics monitoring data to determine multiple frames of steel logistics images.

[0141] Optionally, when the steel logistics image acquisition module 1105 pre-processes the steel logistics monitoring data to determine multiple frames of steel logistics images, it can specifically extract multiple first images to be processed from the steel logistics monitoring data; decode the multiple first images to be processed according to the equipment performance of the monitoring equipment corresponding to the steel logistics monitoring data, and convert the decoded multiple first images to be processed into second images to be processed in the required format; perform image enhancement operations on the second images to be processed to determine multiple frames of steel logistics images.

[0142] Optionally, the data determination module 1101 can specifically search for a pre-trained optional detection model to determine the latest version of the optional detection model as the target detection model; input multiple frames of steel logistics images into the target detection model to determine object detection data; wherein the object detection data includes the position coordinates and name labels of the target detection object in the multiple frames of steel logistics images.

[0143] Optionally, the feature determination module 1102 can specifically select multiple frames of local images corresponding to the position coordinates from the corresponding multiple frames of steel logistics images; analyze the multiple frames of local images to determine the target feature information of the target detection object; wherein the target feature information includes appearance features and motion features.

[0144] Optionally, the tracking result determination module 1103 can specifically obtain a pre-trained target tracking model; input the target feature information into the target tracking model to determine multiple frames of optional tracking images containing the target detection object; calculate the similarity of the target detection object in the multiple frames of optional tracking images based on the target feature information and a preset multi-target tracking algorithm; based on the similarity, eliminate the images corresponding to the abnormal movement trajectory from the multiple frames of optional tracking images to determine the target movement speed of the multiple frames of target tracking images and the target detection object; based on the multiple frames of target tracking images, set a unique code for the target detection object; and generate a target object tracking result based on the unique code, object detection data and target movement speed.

[0145] Optionally, the tracking result determination module 1103 may also perform data formatting on the target object tracking result; and perform tracking and annotation on multiple frames of steel logistics images according to the formatted target object tracking result to generate multiple frames of annotated event management images.

[0146] Optionally, the event information storage module 1104 may specifically receive and analyze event management requirements to obtain corresponding new event results; and determine the new event results, target object tracking results, and multiple-frame annotated event management images as event management result information.

[0147] Optionally, the event management device 1100 for steel logistics further includes an information storage module 1106. The information storage module 1106 can send the multi-frame annotated event management image to the first database for storage through the HTTP protocol, and store the storage path of the multi-frame annotated event management image; and send the newly added event result and the target object tracking result to the second database for storage through the MQTT protocol.

[0148] Optionally, when storing event management result information, the information storage module 1106 can also merge the storage path of multiple-frame annotated event management images with the newly added event results to generate complete event information; store the complete event information in the second database and generate corresponding event records.

[0149] The device performs functions similar to the method provided above. For other functions, please refer to the previous description and will not be repeated here.

[0150] Fig.12 A schematic diagram of the structure of an electronic device according to an exemplary embodiment is shown. Fig.12 As shown, the electronic device 1200 of this embodiment may include: a memory 1201 and a processor 1202 .

[0151] The memory 1201 stores a computer program. When the computer program is executed by the processor 1202, the processor 1202 executes the method in the above embodiment.

[0152] The processor 1202 and the memory 1201 are connected, for example, via a bus.

[0153] Optionally, the electronic device 1200 may further include a transceiver. It should be noted that in actual applications, the number of transceivers is not limited to one, and the structure of the electronic device 1200 does not constitute a limitation on the embodiments of the present application.

[0154] Processor 1202 may be a CPU (Central Processing Unit), a general purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. It may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of this application. Processor 1202 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0155] The bus may include a path to transmit information between the above components. The bus may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0156] The memory 1201 can be a ROM (Read Only Memory) or other types of static storage devices that can store static information and instructions, a RAM (Random Access Memory) or other types of dynamic storage devices that can store information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, optical disk storage (including compressed optical disk, laser disk, optical disk, digital versatile disk, Blu-ray disk, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited to this.

[0157] The memory 1201 is used to store the application code for executing the solution of the present application, and the execution is controlled by the processor 1202. The processor 1202 is used to execute the application code stored in the memory 1201 to implement the contents shown in the above method embodiment.

[0158] The electronic devices include, but are not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), and fixed terminals such as digital TVs, desktop computers, etc. It can also be a server, etc. Fig.12 The electronic 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.

[0159] The electronic device of this embodiment can be used to execute the method of any of the above embodiments, and its implementation principle and technical effect are similar, which will not be described in detail here.

[0160] The present application also provides a non-transitory computer-readable storage medium having computer-readable instructions stored thereon. When the aforementioned instructions are executed by a processor, the processor executes the method in the above embodiment.

[0161] Those skilled in the art can understand that all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to program instructions. The aforementioned program can be stored in a non-transient computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, disk or optical disk and other media that can store program codes.

[0162] The embodiments of the present application are described in detail above. Specific examples are used herein to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method and its core idea of ​​the present application. At the same time, changes or deformations made by those skilled in the art based on the ideas of the present application, the specific implementation methods and the scope of application of the present application, all belong to the scope of protection of the present application. In summary, the content of this specification should not be construed as a limitation on the present application.

Claims

1. An event management method for steel logistics, characterized in that: include: Determining object detection data corresponding to a target detection object in the steel logistics image according to multiple frames of steel logistics images and a pre-trained target detection model; Performing image analysis on the corresponding multiple frames of steel logistics images based on the object detection data to determine target feature information corresponding to the target detection object; Generate a target object tracking result based on the target feature information, the object detection data and the multiple frames of steel logistics images; Event management result information is generated according to the event management requirements and the target object tracking results.

2. The method according to claim 1, characterized in that Before determining the object detection data corresponding to the target detection object in the steel logistics image according to the multiple frames of steel logistics images and the pre-trained target detection model, the method further includes: Receiving steel logistics monitoring data; wherein the steel logistics monitoring data is detected and sent by monitoring equipment installed in a steel logistics site; The steel logistics monitoring data is pre-processed to obtain the multiple frames of steel logistics images.

3. The method according to claim 2, characterized in that The pre-processing of the steel logistics monitoring data to obtain the multiple frames of steel logistics images includes: Extracting a plurality of first images to be processed from the steel logistics monitoring data; Decoding the plurality of first images to be processed according to the equipment performance of the monitoring equipment corresponding to the steel logistics monitoring data, and converting the decoded plurality of first images to be processed into second images to be processed in a required format; An image enhancement operation is performed on the second image to be processed to obtain the multiple frames of steel logistics images.

4. The method according to claim 1, characterized in that: The determining, based on the multiple frames of steel logistics images and the pre-trained target detection model, the object detection data corresponding to the target detection object in the steel logistics image comprises: Searching for a pre-trained optional detection model to determine the latest version of the optional detection model as the target detection model; The multiple frames of steel logistics images are input into the target detection model to determine the object detection data; wherein the object detection data includes the position coordinates and name labels of the target detection objects in the multiple frames of steel logistics images.

5. The method according to claim 4, characterized in that The performing image analysis on the corresponding multiple frames of steel logistics images based on the object detection data to determine target feature information corresponding to the target detection object includes: Selecting multiple frames of local images corresponding to the position coordinates from the corresponding multiple frames of steel logistics images; Analyze the multiple frames of local images to determine target feature information of the target detection object; wherein the target feature information includes appearance features and motion features.

6. The method according to claim 1, characterized in that The generating of the target object tracking result based on the target feature information, the object detection data and the multiple frames of steel logistics images includes: Get a pre-trained target tracking model; Inputting the target feature information into the target tracking model to determine multiple frames of optional tracking images containing the target detection object; Calculating the similarity of the target detection objects in the multiple frames of optional tracking images according to the target feature information and a preset multi-target tracking algorithm; According to the similarity, images corresponding to abnormal movement trajectories are eliminated from the multiple frames of optional tracking images to determine the target movement speed of the multiple frames of target tracking images and the target detection object; According to the multiple frames of target tracking images, a unique code is set for the target detection object; The target object tracking result is generated according to the unique code, the object detection data and the target moving speed.

7. The method according to claim 6, characterized in that The generating of the target object tracking result based on the target feature information, the object detection data and the multiple frames of steel logistics images also includes: Formatting data of the target object tracking result; According to the formatted target object tracking result, the multiple frames of steel logistics images are tracked and annotated to generate multiple frames of annotated event management images.

8. The method according to claim 7, characterized in that The generating of event management result information according to the event management requirements and the target object tracking result includes: Receive the event management requirements and analyze them to obtain corresponding new event results; The newly added event result, the target object tracking result and the multiple frames of annotated event management images are determined as the event management result information.

9. The method according to claim 8, characterized in that The method further includes storing the event management result information, which includes: Sending the multi-frame annotated event management image to a first database for storage through the HTTP protocol, and storing a storage path of the multi-frame annotated event management image; Through the MQTT protocol, the newly added event result and the target object tracking result are sent to the second database for storage.

10. The method according to claim 9, characterized in that The storing of the event management result information further includes: Merging the storage path of the multi-frame annotated event management image with the newly added event result to generate complete event information; The complete event information is stored in the second database, and a corresponding event record is generated.

11. An event management device for steel logistics, characterized in that: include: A data determination module, used to determine object detection data corresponding to a target detection object in the steel logistics image according to multiple frames of steel logistics images and a pre-trained target detection model; A feature determination module, configured to perform image analysis on the corresponding multiple frames of steel logistics images based on the object detection data to determine target feature information corresponding to the target detection object; A tracking result determination module, used for generating a target object tracking result based on the target feature information, the object detection data and the multiple frames of steel logistics images; The event information generation module is used to generate event management result information according to event management requirements and the target object tracking results.

12. An electronic device, characterized in that: include: processor; A memory storing a computer program, which, when executed by the processor, enables the processor to execute the event management method for steel logistics according to any one of claims 1 to 10.

13. A non-transitory computer-readable storage medium, characterized in that: Computer-readable instructions are stored thereon, and when the instructions are executed by a processor, the processor executes the event management method for steel logistics as described in any one of claims 1-10.