A track video analysis method, apparatus, device, medium and product

Abstract

This invention discloses a method, apparatus, equipment, medium, and product for track video analysis, relating to the field of rail transit technology. The method involves acquiring a target analysis task issued by a scheduling platform; determining target analysis nodes and target analysis algorithms based on the target analysis task and an algorithm library; the target analysis node is determined from at least one candidate analysis node at the station node layer based on the target analysis task; the target analysis algorithm is then distributed to the target analysis node, enabling the target analysis node to analyze the video stream based on the algorithm and obtain analysis results; the video stream is a station video stream collected in real time by the candidate analysis nodes. This technical solution, by scheduling the station video stream analysis task in real time through the network layer, helps to achieve reasonable resource allocation and centralized management, solves the problem of low processing speed in existing video data, and enables timely generation of analysis results, ensuring the safe operation of urban rail transit.

Description

Technical Field

[0001] This invention relates to the field of rail transit technology, and in particular to a method, apparatus, equipment, medium, and product for rail video analysis. Background Technology

[0002] With the rapid advancement of urbanization, urban rail transit, with its large capacity, high speed, and high punctuality, has gradually become an important way to solve urban traffic congestion and improve travel efficiency. The expansion of the rail transit network and the increase in operational complexity have made ensuring passenger safety, accelerating emergency response speed, and improving operational efficiency the core challenges for urban rail transit operators.

[0003] Traditional urban rail video surveillance systems rely primarily on manual monitoring, which is not only time-consuming and labor-intensive but also susceptible to subjective factors, resulting in poor monitoring effectiveness. Furthermore, in terms of data processing speed, traditional video analysis systems are limited by hardware performance and algorithm efficiency, making it difficult to handle large-scale video data streams, leading to slow analysis processes and difficulty in responding to various emergencies in a timely manner. Moreover, due to the inability to share front-end resources, traditional video analysis systems need to be constructed in phases according to different lines and different stages, resulting in weak system scalability and upgrade capabilities. Older systems are difficult to adapt to new technical requirements and business needs, leading to short system lifecycles and high maintenance costs.

[0004] Therefore, there is an urgent need for a track video analysis method to achieve real-time analysis, generate alarms and analysis results in a timely manner, and ensure the safe operation of urban rail transit. Summary of the Invention

[0005] This invention provides a method, apparatus, equipment, medium, and product for track video analysis, which solves the problem of low processing speed of existing video data. By adopting this technical solution and setting up a layered architecture, the analysis tasks of station video streams are scheduled in real time through the network layer, which helps to achieve reasonable allocation and centralized management of resources, and can generate analysis results in a timely manner, thus ensuring the safe operation of urban rail transit.

[0006] According to one aspect of the present invention, a track video analysis method is provided, applied to a mesh layer, comprising:

[0007] Obtain the target analysis task issued by the scheduling platform;

[0008] The target analysis nodes and target analysis algorithms are determined based on the target analysis task and algorithm library; the target analysis nodes are determined from at least one candidate analysis node at the station node layer based on the target analysis task.

[0009] The target analysis algorithm is distributed to the target analysis node so that the target analysis node can analyze the video stream based on the target analysis algorithm and obtain the analysis results; the video stream is the station video stream collected in real time by the candidate analysis node.

[0010] According to another aspect of the present invention, a track video analysis method is provided, applied to a target analysis node, comprising:

[0011] Obtain the target analysis task and target analysis algorithm issued by the wire mesh layer;

[0012] Video stream determination is based on target analysis tasks;

[0013] The video stream is converted to obtain at least one image frame;

[0014] Anomaly analysis is performed on the image frame based on the target analysis algorithm to obtain the analysis results.

[0015] According to another aspect of the present invention, a track video analysis apparatus is provided, applied to a mesh layer, comprising:

[0016] The acquisition module is used to acquire target analysis tasks issued by the scheduling platform;

[0017] The determination module is used to determine the target analysis node and the target analysis algorithm based on the target analysis task and the algorithm library; the target analysis node is determined from at least one candidate analysis node in the station node layer based on the target analysis task;

[0018] The analysis module is used to distribute the target analysis algorithm to the target analysis node, so that the target analysis node can analyze the video stream based on the target analysis algorithm and obtain the analysis results; the video stream is a station video stream collected in real time by the candidate analysis node.

[0019] According to another aspect of the present invention, a track video analysis apparatus is provided, applied to a target analysis node, comprising:

[0020] The acquisition module is used to acquire the target analysis tasks and target analysis algorithms issued by the wire mesh layer.

[0021] The determination module is used to determine the video stream based on the target analysis task.

[0022] The conversion module is used to convert the video stream to obtain at least one image frame.

[0023] The analysis module is used to perform anomaly analysis on image frames based on target analysis algorithms and obtain analysis results.

[0024] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0025] At least one processor; and

[0026] A memory communicatively connected to the at least one processor; wherein,

[0027] The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the track video analysis method according to any embodiment of the present invention.

[0028] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the track video analysis method according to any embodiment of the present invention.

[0029] According to another aspect of the present invention, a computer program product is provided, the computer program product comprising a computer program that, when executed by a processor, implements the track video analysis method according to any embodiment of the present invention.

[0030] The technical solution of this invention, by setting up a layered architecture and scheduling the analysis tasks of station video streams in real time through the network layer, helps to achieve reasonable allocation and centralized management of resources, solves the problem of low processing speed of existing video data, and can generate analysis results in a timely manner, thus ensuring the safe operation of urban rail transit.

[0031] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0033] Figure 1 This is a flowchart of a track video analysis method provided according to an embodiment of the present invention;

[0034] Figure 2 This is a flowchart of a track video analysis method provided according to an embodiment of the present invention;

[0035] Figure 3 This is a flowchart of a target analysis node determination method provided according to an embodiment of the present invention;

[0036] Figure 4 This is a flowchart of a track video analysis method provided according to an embodiment of the present invention;

[0037] Figure 5 This is a design drawing of a track video analysis system provided according to an embodiment of the present invention;

[0038] Figure 6 This is a schematic diagram of the structure of a track video analysis device according to an embodiment of the present invention;

[0039] Figure 7 This is a schematic diagram of the structure of a track video analysis device according to an embodiment of the present invention;

[0040] Figure 8 This is a schematic diagram of the structure of an electronic device that implements the track video analysis method of this invention. Detailed Implementation

[0041] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0042] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0043] Furthermore, it should be noted that the information collected in the technical solution of this invention is information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of related data all comply with the relevant laws, regulations and standards of relevant countries and regions, necessary confidentiality measures have been taken, and public order and good morals are not violated. Corresponding operation entry points are provided for users to choose to authorize or refuse.

[0044] Figure 1 This invention provides a flowchart of a track video analysis method. This invention is applicable to analyzing station video streams collected from urban rail transit systems, particularly when the computational resources required for station video streams are excessive. The method can be executed by a track video analysis device, which can be implemented in hardware and / or software and can be configured within the network layer. Figure 1 As shown, the method includes:

[0045] S110. Obtain the target analysis task issued by the scheduling platform.

[0046] The scheduling platform can be a business application system; it issues video analysis tasks and obtains video analysis results; the target analysis task is an instruction to analyze the video data collected in real time at the station, which may include instructions to perform real-time video analysis on the target station, or instructions to analyze historical video data collected from other station nodes.

[0047] Specifically, it involves obtaining and executing intelligent analysis tasks on station video data from the dispatching platform.

[0048] S120. Determine the target analysis nodes and target analysis algorithms based on the target analysis task and algorithm library; the target analysis nodes are determined from at least one candidate analysis node in the station node layer based on the target analysis task.

[0049] The algorithm library stores various algorithm package data and provides corresponding algorithm resources according to the needs of the network layer or station layer; the candidate analysis node is an analysis node corresponding to each station in the station node layer; the target analysis node is the station node that performs the target analysis task; the target analysis algorithm can be at least one algorithm package, which is used as the algorithm data for performing the target analysis task.

[0050] Specifically, the target analysis node is determined from at least one candidate analysis node in the station node layer based on the target analysis task, and the target analysis algorithm is determined based on the target analysis task and the algorithm library.

[0051] Understandably, by configuring the algorithm library and distributing the target algorithm, unified management and flexible deployment of algorithms are achieved, enabling them to adapt to ever-changing business task requirements.

[0052] S130. The target analysis algorithm is sent to the target analysis node so that the target analysis node can analyze the video stream based on the target analysis algorithm and obtain the analysis results; the video stream is the station video stream collected in real time by the candidate analysis node.

[0053] The video stream can be a video stream collected in real time by any candidate analysis node at the station node layer through cameras deployed at the station; it can be a real-time video stream or a historical video stream; the historical video stream can be a video stream collected during non-operational periods or during peak data processing periods; the analysis can be image recognition or abnormal behavior recognition of the video stream by the target analysis node through target analysis algorithms; the analysis result can be an image or abnormal scene corresponding to the video stream with abnormal behavior, such as abnormal results such as construction pile foundation monitoring, monitoring of not wearing safety helmets, irrelevant behavior at the construction site, and abnormal intrusion monitoring in the construction scene; or abnormal results such as on-duty personnel sleeping or absent from duty in the station scene, garbage not cleaned in public areas, signs not lit, and abnormal equipment status in the station scene.

[0054] Specifically, the target analysis algorithm is sent to the video analysis server on the target analysis node in the station layer, and video analysis is performed on the video stream corresponding to the target task to generate preliminary analysis results, which may include abnormal scenes, abnormal behaviors in abnormal scenes, and the image frames to which the abnormal behaviors belong.

[0055] In one optional embodiment of the present invention, after obtaining the analysis results, the analysis results are filtered and optimized, and the analysis results are used to issue early warning prompts through the scheduling platform: after optimizing and filtering the preliminary results, the network layer feeds back the video analysis result data to the business application system through an interactive interface or pop-up warning; this further improves data quality, ensures that the network layer can respond quickly to the on-site situation, generate alarms and analysis results in a timely manner, and ensure the safe operation of urban rail transit.

[0056] This invention, in its embodiments, acquires target analysis tasks issued by a scheduling platform; determines target analysis nodes and target analysis algorithms based on the target analysis tasks and an algorithm library; the target analysis nodes are determined from at least one candidate analysis node at the station node layer based on the target analysis tasks; the target analysis algorithms are then distributed to the target analysis nodes, enabling them to analyze the video stream based on the algorithms and obtain analysis results; the video stream is the station video stream collected in real time by the candidate analysis nodes. This technical solution, by scheduling the station video stream analysis tasks in real time through the network layer, helps to achieve reasonable resource allocation and centralized management, solves the problem of low processing speed in existing video data, and enables timely generation of analysis results, thus ensuring the safe operation of urban rail transit.

[0057] Figure 2 This is a flowchart of a track video analysis method according to an embodiment of the present invention. Based on the above embodiments, this embodiment supplements the determination method of the target analysis node and the target analysis algorithm. It should be noted that for parts not described in detail in this embodiment, please refer to the relevant descriptions in other embodiments. Figure 2As shown, the method includes:

[0058] S210. Obtain the target analysis task issued by the scheduling platform.

[0059] S220. Analyze the target analysis task to obtain the task type, task node identifier, and algorithm request.

[0060] The task type can be the task requirements of the target analysis task, such as analyzing the real-time video stream collected by the task node, or analyzing the historical video stream collected by each station in the station node layer; the task node identifier is the identifier of the analysis node specified by the target analysis task; and the algorithm request is the algorithm required to execute the target analysis task.

[0061] Specifically, the target analysis task is parsed to obtain the task type, task node identifier, and algorithm request of the target analysis task.

[0062] S230. Determine the target analysis algorithm from the algorithm library based on the algorithm request.

[0063] The target analysis algorithm is stored as an algorithm package data.

[0064] Specifically, based on the algorithm request, at least one algorithm package data required to perform the target analysis task is found from the algorithm library.

[0065] S240. Determine the target analysis node from at least one candidate analysis node in the station node layer based on the task type and task node identifier.

[0066] The task type can be a real-time task or a historical task; historical tasks can be historical video stream analysis tasks stored during peak video analysis periods, or historical video stream analysis tasks that other nodes cannot process, or historical video stream analysis tasks during non-operational periods of the station.

[0067] Specifically, the task type is determined. If the task type is a real-time task, the target analysis node is determined from at least one candidate analysis node in the station node layer based on the task node identifier. If the task type is a historical task, it is a specified execution, and the corresponding analysis node is directly determined as the target analysis node based on the task node identifier.

[0068] Optional, such as Figure 3 The method for determining a target analysis node, as shown, determines the target analysis node from at least one candidate analysis node in the station node layer based on the task type and task node identifier, including:

[0069] S241. Determine the task type. If the task type is a historical task, determine the target analysis node from the station node layer based on the task node identifier.

[0070] Specifically, if the task type is a historical task, the instruction for the target analysis task is to analyze the historical video stream stored during peak periods, or to analyze the historical video stream that other nodes cannot process, or to analyze the historical video stream during non-operational periods of the station; if it is a designated analysis node, the target analysis node to which the task node identifier belongs is directly determined from the station node layer based on the task node identifier.

[0071] S242. If the task type is a real-time task, the target analysis node is determined from at least one candidate analysis node in the station node layer based on the task node identifier, the target analysis task, and the resource data threshold.

[0072] Among them, the real-time task is an instruction to analyze the video stream collected in real time by the analysis node to which the task node identifier belongs; the target analysis task may include video stream information, such as the amount of resources used; the resource data threshold is the resource threshold that the task node identifier can use to execute this analysis task.

[0073] Specifically, if the task type is a real-time task, the analysis node is determined based on the task node identifier, and the target analysis node is determined from at least one candidate analysis node in the station node layer based on the analysis node, the target analysis task, and the resource data threshold.

[0074] Optionally, the target analysis node is determined from at least one candidate analysis node in the station node layer based on the task node identifier, the target analysis task, and resource data thresholds, including:

[0075] The initial analysis node is determined from the station node layer based on the task node identifier, and the available resources of the initial analysis node are obtained.

[0076] The available resources of a node are compared with the resource data threshold to obtain the first comparison result;

[0077] If the first comparison result is that the available resources of a node are greater than the resource data threshold, then the initial analysis node will be used as the target analysis node.

[0078] If the first comparison result is that the available resources of a node are less than the resource data threshold, then the video stream information is determined based on the target analysis task;

[0079] The target analysis node is determined from at least one candidate analysis node in the station node layer based on the analysis level and the amount of analysis data of the video stream information.

[0080] The initial analysis node is the analysis node corresponding to the task node identifier, i.e., the corresponding station; the available resources of the node are the amount of resources that the analysis node can use to perform video analysis; the analysis level of the video stream information is the importance of the target analysis task; a high level indicates that the target analysis task is important and needs to be analyzed immediately; a low level indicates that the target analysis task is not important and can be analyzed later; the analysis data volume of the video stream information is the amount of data used to analyze the video stream, i.e., the amount of resources used for analysis.

[0081] Specifically, based on the task node identifier in this target analysis task, the corresponding initial analysis node is determined from the station node layer, and the current available resource quantity of the initial analysis node is further obtained from the station node layer. The current available resource quantity of the initial analysis node is compared with the resource data threshold. If the available resource quantity is greater than the resource data threshold, it indicates that the initial analysis node can execute the target analysis task, and the initial analysis node is used as the target analysis node for target task analysis. If the available resource quantity is less than the resource data threshold, it indicates that the initial analysis node cannot support this target analysis task. The video stream information corresponding to the target analysis task is further determined, and the target analysis node is determined from at least one candidate analysis node in the station node layer to execute this target analysis task based on the analysis level and analysis data volume of the video stream information.

[0082] Understandably, by enabling local and timely analysis through the initial analysis node, the system can rationally allocate and schedule target analysis tasks based on available resources, avoiding resource waste and further improving the overall performance of the system.

[0083] Optionally, the target analysis node is determined from at least one candidate analysis node in the station node layer based on the analysis level and analysis data volume of the video stream information, including:

[0084] If the analysis level of the video stream information is high, then the target analysis node is determined from at least one candidate analysis node based on the amount of analysis data and the available resource data of at least one candidate analysis node in the station node layer.

[0085] If the analysis level of the video stream information is low, the target analysis node is determined from at least one candidate analysis node based on the amount of analysis data and the scheduling data threshold.

[0086] Among them, the available resource data refers to the amount of available resources of each candidate analysis node in the station node layer; the scheduling data threshold is whether the video stream information meets the redistribution of the analysis task.

[0087] Specifically, if the analysis level of the video stream information is high, it indicates that the target analysis task needs to be analyzed immediately. In this case, the available resource data of the other candidate analysis nodes in the station node layer is obtained, and the candidate analysis node with available resource data greater than the analysis data volume of the target analysis task is found. This candidate analysis node is then used as the target analysis node for executing the target analysis task. If the analysis level of the video stream information is low, it indicates that its execution can be delayed. In this case, it is determined whether reallocation can be carried out based on the analysis data volume and the scheduling data threshold.

[0088] Understandably, by analyzing video stream information at different levels and setting scheduling data thresholds, while ensuring resource allocation, timely analysis of important analytical tasks can be effectively guaranteed, front-end resource sharing can be achieved, the adaptability of the intelligent video analysis system can be improved, it is not limited by hardware performance, it can adapt to new technical requirements and business needs, and reduce later maintenance costs.

[0089] Optionally, the target analysis node is determined from at least one candidate analysis node based on the amount of analysis data and the scheduling data threshold, including:

[0090] By comparing the amount of data analyzed with the scheduling data threshold, a second comparison result is obtained;

[0091] If the second comparison result is that the amount of data to be analyzed is less than the scheduling data threshold, then the initial analysis node will be used as the target analysis node.

[0092] If the second comparison result is that the amount of analysis data is greater than the scheduling data threshold, then obtain the available resource data of at least one candidate analysis node;

[0093] Based on the amount of data to be analyzed and the available resource data, the target analysis node is determined from at least one candidate analysis node.

[0094] Specifically, the analysis data volume is compared with the scheduling data threshold. If the analysis data volume is less than the scheduling data threshold, it indicates that the current target analysis task does not meet the conditions for task reallocation. The initial analysis node is then used as the target analysis node, and the execution of the analysis task is delayed. If the analysis data volume is greater than the scheduling data threshold, it indicates that the current target analysis task meets the conditions for task reallocation. In this case, the available resource data of at least one candidate analysis node at the station node layer is obtained. Based on the analysis data volume and available resource data, a candidate analysis node with an analysis data volume less than the available resource data is found, and this candidate analysis node is used as the target analysis node.

[0095] In one optional embodiment of the present invention, when the amount of analysis data is less than the scheduling data threshold, i.e., the target analysis task does not meet the conditions for task reallocation, the target analysis task can be deferred and treated as a historical analysis task. A deferred interval is set for the historical analysis task. After the deferred analysis time is reached based on the current timestamp and the deferred interval, the initial analysis node performs video intelligent analysis on the historical analysis task. This ensures that the analysis of less important tasks is deferred, achieves reasonable allocation of intelligent analysis tasks, and does not occupy the available resources of other candidate analysis nodes.

[0096] In an optional embodiment of the present invention, when the amount of analysis data is less than the scheduling data threshold, i.e., the target analysis task does not meet the conditions for task reallocation, the available resources of the network layer can also be used for judgment. If the available resources of the network layer meet the amount of analysis data, then the network layer performs intelligent video analysis, realizing the overall allocation of task computing resources. Based on the available resources, other analysis nodes in the network layer or station node layer perform synchronous analysis, further improving data processing efficiency. It can be understood that by setting a scheduling data threshold to realize the reallocation of target analysis tasks, it is in line with the characteristics of video surveillance systems with large storage capacity and high pressure for real-time parsing of backhaul data. This helps to achieve reasonable allocation and centralized management of resources, and the local processing capability of the station node layer enables the system to respond quickly to the situation on site, generate alarms and analysis results in a timely manner, and ensure the safe operation of urban rail transit.

[0097] S250. The target analysis algorithm is sent to the target analysis node so that the target analysis node can analyze the video stream based on the target analysis algorithm and obtain the analysis results; the video stream is the station video stream collected in real time by the candidate analysis node.

[0098] This invention provides a method for rationally allocating and scheduling target analysis tasks based on task type. It supports flexible access to intelligent analysis servers at the network layer or station node layer, enabling the overall allocation and management of computing resources. Tasks can be assigned to other stations at the network layer or station node layer as needed, thus achieving a rational allocation of computing resources, avoiding resource waste, and improving the overall performance of the system. Furthermore, based on task sharing allocation and scheduling, it effectively improves the reusability of front-end resources, ensuring that each candidate analysis node is applicable to different lines. After the phased monitoring target is completed, the computing power can be released and applied to subsequent analysis tasks.

[0099] Figure 4This invention provides a flowchart of a track video analysis method. This invention is applicable to analyzing station video streams collected from urban rail transit systems, particularly when station video streams consume excessive computing resources. The method can be executed by a track video analysis device, which can be implemented in hardware and / or software and configured within the target analysis node. Figure 4 As shown, the method includes:

[0100] S310. Obtain the target analysis task and target analysis algorithm issued by the wire mesh layer.

[0101] Specifically, the target analysis algorithm determined by the target analysis task is obtained from the wire mesh layer, and the target analysis instructions for this task are further obtained.

[0102] S320. Determine the video stream based on the target analysis task.

[0103] The video stream can be a station video stream collected by the camera deployed on the target analysis node; it can also be a station video stream collected historically by the target analysis node; or it can be a station video stream collected by other candidate analysis nodes in the station node layer.

[0104] Specifically, the station video stream corresponding to this analysis instruction is determined based on the target analysis task.

[0105] S330. Convert the video stream to obtain at least one image frame.

[0106] Here, an image frame is a sequence of image frames corresponding to a video stream.

[0107] Specifically, the video stream is transformed to obtain at least one image frame corresponding to the video stream of the station to be analyzed.

[0108] S340. Anomaly analysis is performed on the image frame based on the target analysis algorithm to obtain the analysis results.

[0109] Anomaly analysis can be used for image scene recognition and behavior recognition; it can also be an analysis method performed by existing artificial intelligence video analysis systems.

[0110] Specifically, convolutional neural network algorithms can be used to identify image scenes in image frames. Furthermore, based on each image scene, techniques such as recurrent neural networks and long short-term memory networks can be used to model and analyze continuous behaviors in the image scene, identify abnormal behaviors of passengers such as falling or fighting, and associate the abnormal behavior, image scene and corresponding image frame as the analysis result.

[0111] In one optional embodiment of the present invention, before performing anomaly analysis on the image frame, the image frame can be preprocessed, such as by using data augmentation techniques such as rotation, scaling, and flipping of the original data, and by using a pre-trained model to perform transfer learning on the image frame with a small amount of labeled data, thereby improving the accuracy and generalization ability of the recognition and achieving high-precision recognition of abnormal behavior.

[0112] This invention, through the deployment of an artificial intelligence video analysis system on analysis nodes at each station node layer, achieves a significant improvement in data processing speed compared to traditional urban rail video analysis systems. This is achieved by utilizing artificial intelligence technologies such as deep learning. Furthermore, by setting up at least one analysis node at the station layer and employing multi-threading technology and asynchronous processing mechanisms, real-time processing of massive amounts of video data is realized, further improving data processing speed and ensuring that efficient data processing capabilities are maintained even when dealing with complex scenarios such as image obstruction and crowded conditions.

[0113] Figure 5 This is a design diagram of a track video analysis system according to an embodiment of the present invention, which includes a safety production network of the main control center and an enterprise management network of the backup control center at the network layer; and each station node at the station node layer.

[0114] The safety production network at the network layer interacts with each station node in the station node layer to perform analysis tasks and algorithms. It is responsible for overall resource management, data optimization, and algorithm scheduling. It includes storage devices, intelligent analysis servers, and general-purpose servers. Storage devices store target analysis tasks and results. Intelligent analysis servers execute a small number of analysis tasks. It also includes a resource management and scheduling platform for unified management and scheduling of the system's computing resources. This platform receives intelligent analysis tasks and submits algorithm loading requests to the algorithm repository based on task requirements. The algorithm repository stores various algorithm package data and provides corresponding algorithm resources according to the needs of the resource management and scheduling platform or the station layer, achieving unified management and sharing of algorithms. The algorithm strategy platform interacts with the data resource pool to optimize and filter analysis results, improving data quality. The data resource pool stores and manages various types of data in the system, providing data support to the algorithm strategy platform and receiving optimized and filtered result data for data recycling and optimization. Algorithm engine edge nodes are responsible for deploying algorithm package data from the algorithm repository to the intelligent analysis servers of each station node in the network layer or station layer, enabling rapid response to local analysis needs and reducing data transmission latency.

[0115] The enterprise management network at the wired layer serves as a backup control center. If the security production network is unable to support algorithm scheduling or the stored data resources are damaged, the enterprise management network can be used as an alternative to perform algorithm resource scheduling or obtain backup data resources from its storage devices.

[0116] At least one station analysis node in the station layer includes a storage device and an intelligent analysis server. The storage device is used to store the collected video streams and analysis results. The intelligent analysis server is used to receive the target analysis tasks and target analysis algorithms issued by the network layer and execute the analysis instructions. In addition, each station in the station layer is equipped with a station camera, which is responsible for collecting the station's video streams in real time to provide raw data for intelligent analysis.

[0117] Intelligent analysis servers can employ artificial intelligence video analysis systems, utilizing computer vision and artificial intelligence technologies to mimic human visual cognitive mechanisms. By preprocessing video streams using computer image visual analysis techniques—such as frame rate control, image enhancement, and noise suppression—video quality is improved, significantly reducing the negative impact of poor front-end input image quality on subsequent intelligent analysis. Furthermore, it can automatically understand and separate background and targets within a scene, thereby analyzing and tracking targets appearing in the scene. Urban rail users can preset different alarm rules for different camera application scenarios based on the artificial intelligence video analysis function. Once a target in the scene exhibits behavior that violates or conforms to the predefined rules, the artificial intelligence video analysis system will automatically issue an alarm. It can accurately detect specific areas, people, and objects, or monitor passenger density and traffic statistics, effectively avoiding the limitations of traditional urban rail transit video analysis systems that lack autonomous learning and adaptive adjustment capabilities, making it difficult to automatically optimize the recognition model according to environmental changes and newly emerging scenarios. Additionally, the recognition accuracy of traditional video analysis systems is not ideal due to factors such as light, angle, and occlusion, especially in complex and ever-changing urban rail transit scenarios, where false alarm and missed alarm rates are high.

[0118] In this embodiment of the invention, by constructing a network layer and a station node layer, the limitations of traditional video analysis systems in handling large-scale video data streams due to hardware performance and algorithm efficiency are avoided, resulting in slow analysis processes and difficulty in responding to various emergencies in a timely manner. By judging the available resources of the station node layer, front-end resources of different lines and different stages of construction are shared, improving the availability of intelligent video analysis. Furthermore, by configuring an artificial intelligence analysis system, the lack of self-learning and adaptive adjustment capabilities of traditional urban rail transit video analysis systems is effectively avoided, making it difficult to automatically optimize the recognition model according to environmental changes and new scenarios. This ensures that the false alarm rate and false negative rate of intelligent video analysis are reduced in complex and ever-changing urban rail transit scenarios.

[0119] Figure 6This invention provides a schematic diagram of a track video analysis device, applicable to the analysis of station video streams collected from urban rail transit, particularly suitable for situations where station video streams consume excessive computing resources. This track video analysis device can be implemented in hardware and / or software and can be configured within the network layer. Figure 6 As shown, the track video analysis device 400 includes an acquisition module 410, a determination module 420, and an analysis module 430:

[0120] Module 410 is used to acquire target analysis tasks issued by the scheduling platform;

[0121] The determination module 420 is used to determine the target analysis nodes and target analysis algorithms based on the target analysis task and the algorithm library; the target analysis nodes are determined from at least one candidate analysis node in the station node layer based on the target analysis task;

[0122] The analysis module 430 is used to distribute the target analysis algorithm to the target analysis node, so that the target analysis node can analyze the video stream based on the target analysis algorithm and obtain the analysis results; the video stream is the station video stream collected in real time by the candidate analysis node.

[0123] This invention, in its embodiments, acquires target analysis tasks issued by a scheduling platform; determines target analysis nodes and target analysis algorithms based on the target analysis tasks and an algorithm library; the target analysis nodes are determined from at least one candidate analysis node at the station node layer based on the target analysis tasks; the target analysis algorithms are then distributed to the target analysis nodes, enabling them to analyze the video stream based on the algorithms and obtain analysis results; the video stream is the station video stream collected in real time by the candidate analysis nodes. This technical solution, by scheduling the station video stream analysis tasks in real time through the network layer, helps to achieve reasonable resource allocation and centralized management, solves the problem of low processing speed in existing video data, and enables timely generation of analysis results, thus ensuring the safe operation of urban rail transit.

[0124] Optionally, the determination module 420 includes a parsing unit, a determination unit, and a target analysis node determination unit;

[0125] The parsing unit is used to parse the target analysis task to obtain the task type, task node identifier, and algorithm request;

[0126] The determination unit is used to determine the target analysis algorithm from the algorithm library based on the algorithm request;

[0127] The target analysis node determination unit is used to determine the target analysis node from at least one candidate analysis node in the station node layer based on the task type and task node identifier.

[0128] Optionally, the target analysis node determination unit may also include a historical task determination unit and a real-time task determination unit;

[0129] The historical task determination unit is used to determine the task type. If the task type is a historical task, the target analysis node is determined from the station node layer based on the task node identifier.

[0130] The real-time task determination unit is used to determine the target analysis node from at least one candidate analysis node in the station node layer based on the task node identifier, the target analysis task, and the resource data threshold if the task type is a real-time task.

[0131] Optionally, the real-time task determination unit is also used to determine the initial analysis node from the station node layer based on the task node identifier, and to obtain the node available resource quantity of the initial analysis node.

[0132] The available resources of a node are compared with the resource data threshold to obtain the first comparison result;

[0133] If the first comparison result is that the available resources of a node are greater than the resource data threshold, then the initial analysis node will be used as the target analysis node.

[0134] If the first comparison result is that the available resources of a node are less than the resource data threshold, then the video stream information is determined based on the target analysis task;

[0135] The target analysis node is determined from at least one candidate analysis node in the station node layer based on the analysis level and the amount of analysis data of the video stream information.

[0136] Optionally, the real-time task determination unit is also used to determine the target analysis node from at least one candidate analysis node based on the amount of analysis data and the available resource data of at least one candidate analysis node in the station node layer if the analysis level of the video stream information is high.

[0137] If the analysis level of the video stream information is low, the target analysis node is determined from at least one candidate analysis node based on the amount of analysis data and the scheduling data threshold.

[0138] Optionally, the real-time task determination unit is also used to compare and analyze the amount of data with the scheduling data threshold to obtain a second comparison result;

[0139] If the second comparison result is that the amount of data to be analyzed is less than the scheduling data threshold, then the initial analysis node will be used as the target analysis node.

[0140] If the second comparison result is that the amount of analysis data is greater than the scheduling data threshold, then obtain the available resource data of at least one candidate analysis node;

[0141] Based on the amount of data to be analyzed and the available resource data, the target analysis node is determined from at least one candidate analysis node.

[0142] Figure 7 This invention provides a schematic diagram of a track video analysis device. This invention is applicable to analyzing station video streams collected from urban rail transit systems, particularly where station video streams consume excessive computing resources. The track video analysis device can be implemented in hardware and / or software and can be configured within the target analysis node. Figure 7 As shown, the track video analysis device 500 includes an acquisition module 510, a determination module 520, a conversion module 530, and an analysis module 540.

[0143] The acquisition module 510 is used to acquire the target analysis task and target analysis algorithm issued by the wire mesh layer.

[0144] The determination module 520 is used to determine the video stream based on the target analysis task.

[0145] The conversion module 530 is used to convert the video stream to obtain at least one image frame.

[0146] Analysis module 540 is used to perform anomaly analysis on image frames based on target analysis algorithms and obtain analysis results.

[0147] This invention, through the deployment of an artificial intelligence video analysis system on analysis nodes at each station node layer, achieves a significant improvement in data processing speed compared to traditional urban rail video analysis systems. This is achieved by utilizing artificial intelligence technologies such as deep learning. Furthermore, by setting up at least one analysis node at the station layer and employing multi-threading technology and asynchronous processing mechanisms, real-time processing of massive amounts of video data is realized, further improving data processing speed and ensuring efficient data processing capabilities even when dealing with complex scenarios such as image obstruction and crowded conditions.

[0148] The track video analysis device provided in this embodiment of the invention can execute the track video analysis method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0149] According to embodiments of the present invention, the present invention also provides an electronic device, a readable storage medium, and a computer program product.

[0150] Figure 8A schematic diagram of an electronic device 10 that can be used to implement embodiments of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0151] like Figure 8 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0152] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0153] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as track video analysis methods.

[0154] In some embodiments, the track video analysis method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the track video analysis method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the track video analysis method by any other suitable means (e.g., by means of firmware).

[0155] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations may include: implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0156] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0157] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0158] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0159] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0160] A computing system can include clients and servers. Clients and servers are generally geographically separated and typically interact via communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a host product within the cloud computing service system. This addresses the shortcomings of traditional physical hosts and dedicated virtual services, such as high management difficulty and weak business scalability.

[0161] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0162] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for analyzing track video, characterized in that, Applied to the wireframe layer, including: Obtain the target analysis task issued by the scheduling platform; The target analysis nodes and target analysis algorithms are determined based on the target analysis task and algorithm library; the target analysis nodes are determined from at least one candidate analysis node at the station node layer based on the target analysis task. The target analysis algorithm is distributed to the target analysis node so that the target analysis node can analyze the video stream based on the target analysis algorithm and obtain the analysis results; the video stream is the station video stream collected in real time by the candidate analysis node.

2. The method according to claim 1, characterized in that, The process of determining the target analysis node and target analysis algorithm based on the target analysis task and algorithm library includes: The target analysis task is parsed to obtain the task type, task node identifier, and algorithm request; Based on the algorithm request, the target analysis algorithm is determined from the algorithm library; The target analysis node is determined from at least one candidate analysis node in the station node layer based on the task type and the task node identifier.

3. The method according to claim 2, characterized in that, The step of determining the target analysis node from at least one candidate analysis node in the station node layer based on the task type and the task node identifier includes: The task type is determined. If the task type is a historical task, the target analysis node is determined from the station node layer based on the task node identifier. If the task type is a real-time task, then the target analysis node is determined from at least one candidate analysis node in the station node layer based on the task node identifier, the target analysis task, and the resource data threshold.

4. The method according to claim 3, characterized in that, Determining the target analysis node from at least one candidate analysis node in the station node layer based on the task node identifier, the target analysis task, and resource data thresholds includes: Based on the task node identifier, the initial analysis node is determined from the station node layer, and the available node resources of the initial analysis node are obtained. The available resources of the node are compared with the resource data threshold to obtain a first comparison result; If the first comparison result indicates that the available resources of the node are greater than the resource data threshold, then the initial analysis node will be used as the target analysis node. If the first comparison result indicates that the available resources of the node are less than the resource data threshold, then the video stream information is determined based on the target analysis task; The target analysis node is determined from at least one candidate analysis node in the station node layer based on the analysis level and the amount of analysis data of the video stream information.

5. The method according to claim 4, characterized in that, The determination of the target analysis node from at least one candidate analysis node in the station node layer based on the analysis level and analysis data volume of the video stream information includes: If the analysis level of the video stream information is high, then the target analysis node is determined from the at least one candidate analysis node based on the amount of analysis data and the available resource data of at least one candidate analysis node in the station node layer. If the analysis level of the video stream information is low, then the target analysis node is determined from at least one candidate analysis node based on the analysis data volume and the scheduling data threshold.

6. The method according to claim 5, characterized in that, Determining the target analysis node from at least one candidate analysis node based on the amount of analysis data and the scheduling data threshold includes: A second comparison result is obtained by comparing the amount of analyzed data with the scheduling data threshold. If the second comparison result indicates that the amount of analysis data is less than the scheduling data threshold, then the initial analysis node is taken as the target analysis node. If the second comparison result indicates that the amount of analysis data is greater than the scheduling data threshold, then the available resource data of the at least one candidate analysis node is obtained. Based on the amount of analysis data and the available resource data, the target analysis node is determined from at least one candidate analysis node.

7. A method for analyzing track video, characterized in that, Applied to target analysis nodes, including: Obtain the target analysis task and target analysis algorithm issued by the wire mesh layer; Determine the video stream based on the target analysis task; The video stream is converted to obtain at least one image frame; Anomaly analysis is performed on the image frame based on the target analysis algorithm to obtain the analysis results.

8. A track video analysis device, characterized in that, Applied to the wireframe layer, including: The acquisition module is used to acquire target analysis tasks issued by the scheduling platform; The determination module is used to determine the target analysis node and the target analysis algorithm based on the target analysis task and the algorithm library; the target analysis node is determined from at least one candidate analysis node in the station node layer based on the target analysis task; The analysis module is used to distribute the target analysis algorithm to the target analysis node, so that the target analysis node can analyze the video stream based on the target analysis algorithm and obtain the analysis results; the video stream is a station video stream collected in real time by the candidate analysis node.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the track video analysis method according to any one of claims 1-6 or 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the orbital video analysis method according to any one of claims 1-6 or 7.

11. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the track video analysis method according to any one of claims 1-6 or 7.

Images