A biometric system for complex scenarios

Abstract

This application provides a biometric system for complex scenarios, relating to the field of biometric technology. By setting up multiple cameras that can be deployed in any area of ​​the monitoring site, the system allows for flexible adjustment of the camera's installation position and number based on the actual layout of the complex scenario and the key monitoring areas. This achieves full coverage of the monitoring site without blind spots, reducing reliance on manual inspections and solving the monitoring challenges of multiple and irregular areas in complex scenarios. It is suitable for rodent monitoring needs in various complex scenarios such as warehousing, agriculture, and urban public facilities. By accessing the video streams from multiple cameras through an edge computing device and running an AI visual analysis model locally, the system enables real-time local analysis of the video streams and identification of specific organisms (mice). This eliminates the need to upload all video streams to the cloud, reducing network bandwidth consumption and mitigating transmission latency issues caused by centralized cloud processing.

Description

Technical Field

[0001] This invention relates to the field of biometrics, and more specifically, to a biometrics system for complex scenarios. Background Technology

[0002] In various complex scenarios such as warehousing and logistics, agricultural production, urban public facility management, and food processing, rodents are common pests. Their activities not only cause material losses and facility damage but also spread pathogens, seriously affecting the safety, hygiene, and normal operation of these scenarios. Therefore, accurate detection and effective monitoring of rodents have become an urgent need in various scenarios. Currently, rodent detection and identification mainly rely on traditional manual inspections. Traditional manual inspection methods are time-consuming and labor-intensive, and are affected by factors such as fluctuations in human attention and limitations in the inspection range, resulting in high false negative rates and slow response times. Therefore, we propose an improvement to this problem by developing a biometric identification system for complex scenarios. Summary of the Invention

[0003] This invention provides a biometric system for complex scenarios, comprising: Multiple cameras are deployed at the monitoring site; Edge computing devices are used to access video streams captured by multiple cameras and run AI visual analysis models locally to perform real-time analysis of the video streams, identify specific biological targets, and generate structured event data containing target information. The cloud management server communicates with edge computing devices to perform unified registration, activation, status monitoring, remote task distribution, and data aggregation for edge computing devices, and provides a graphical user interface. The cloud management server includes a visual workflow editing module, which allows users to define the processing logic of video analysis by dragging and dropping components in a graphical manner to generate pipeline configurations, and remotely distribute the pipeline configurations to edge computing devices for execution.

[0004] As a preferred technical solution of this application, the edge computing device includes: The AI ​​inference module is used to load and execute deep learning models for object detection. The pipeline management module is used to receive, load, and execute pipeline configurations issued by the cloud management server to schedule video acquisition, AI inference, and result processing. The communication module is used to establish a two-way communication connection with the cloud management server via the MQTT protocol to receive remote commands and report device heartbeats and structured event data.

[0005] As a preferred technical solution of this application, the pipeline configuration generated by the visualization workflow editing module includes an input source component, an AI processing component, and an output component connected in sequence; the AI ​​processing component is associated with a target detection deep learning model, and the target detection deep learning model is remotely hot-updated through a cloud management server.

[0006] As a preferred technical solution of this application, the cloud management server further includes: The device management module is used to manage the list, details, and geographical location of edge computing devices, and provides a remote command issuance interface. The remote commands include device restart, log collection, video streaming control, and firmware upgrade. The data analysis dashboard module is used to receive the structured event data reported by the edge computing device and display the device operation indicators, biometric target recognition statistical trends and spatial distribution in the form of charts.

[0007] As a preferred technical solution of this application, it also includes a mobile terminal application, which is communicatively connected to a cloud management server and is used for: Receive and display real-time event notifications generated by edge computing devices and forwarded by cloud management servers; View the real-time status, geographic location, and historical operating data of edge computing devices; Request and play live video streams or recorded historical videos from cameras associated with the edge computing device.

[0008] As a preferred technical solution of this application, the mobile terminal application establishes a connection with the cloud management server through the WebSocket protocol to receive the real-time event notification; the types of the real-time event notification include target detection events, target capture events, device failure events, and device status change events.

[0009] As a preferred technical solution of this application, the edge computing device automatically detects and switches between at least two networks, including wired network, wireless local area network and cellular mobile network, and automatically reconnects to the cloud management server after the network failure is recovered.

[0010] As a preferred technical solution in this application, the firmware upgrade package issued by the cloud management server adopts a blue-green deployment method, and the edge computing device automatically rolls back to the software version before the upgrade when the upgrade fails.

[0011] As a preferred technical solution of this application, the specific biological target is rodents, which enables rodent infestation monitoring.

[0012] As a preferred technical solution of this application, the following steps are included: On the cloud management server, the configuration of a video analysis pipeline for a specific biological target can be defined through a visual workflow editing module; The pipeline configuration and related AI models are remotely distributed from the cloud management server to the edge computing device; The edge computing device loads and runs the pipeline configuration to perform real-time AI analysis on the accessed local video stream and identify specific biological targets; Edge computing devices report the identified structured event data to the cloud management server; The cloud management server aggregates and displays event data from edge computing devices and pushes key event notifications to users through mobile terminal applications.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: In the scheme of this application: 1. This application sets up multiple cameras that can be placed in any area of ​​the monitoring site. The installation position and number of cameras can be flexibly adjusted according to the actual layout of the complex scene and the key monitoring areas, so as to achieve full coverage of the monitoring site without blind spots, reduce the reliance on manual inspection, and solve the monitoring problem of multiple areas and irregular areas in complex scenes. It is suitable for rodent monitoring needs in various complex scenes such as warehouses, agriculture, and urban public facilities. 2. By accessing video streams from multiple cameras through edge computing devices and running AI visual analysis models locally, real-time local analysis of video streams and identification of specific organisms (mice) can be achieved without uploading all video streams to the cloud, reducing network bandwidth consumption and mitigating transmission latency issues caused by centralized cloud processing. Attached Figure Description

[0014] Figure 1 A schematic diagram of a biometric system for complex scenarios provided in this application; Figure 2 A schematic diagram of the camera and edge computing device provided in this application; Figure 3 A structural schematic diagram of the camera and edge computing device provided in this application from another perspective; Figure 4 A flowchart of the biometric system for complex scenarios provided in this application. Detailed Implementation

[0015] 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 of the present invention. 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.

[0016] It should be noted that, unless otherwise specified, the embodiments and features and technical solutions in the present invention can be combined with each other.

[0017] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0018] For an example, please refer to... Figures 1-4 A biometric system for complex scenarios, comprising multiple cameras 1, deployed at the monitoring site; by directly deploying multiple cameras 1 at the monitoring site, it is possible to achieve no blind spots in the target area, ensuring full-scene capture of biometric targets; Edge computing device 2 is used to access video streams captured by multiple cameras 1 and run an AI visual analysis model locally to perform real-time analysis of the video streams, identify specific biological targets, and generate structured event data containing target information. Relying on the local computing capabilities of edge computing device 2, real-time analysis and target recognition of video streams are achieved. Because local computing does not require uploading the entire video stream to the cloud, it reduces the bandwidth consumption and latency of data transmission. At the same time, target recognition can be completed locally to quickly generate structured event data, enabling immediate perception and feedback of biological targets and reducing the monitoring lag caused by cloud computing and transmission time. The cloud management server communicates with edge computing device 2 and is used for unified registration, activation, status monitoring, remote task distribution and data aggregation of edge computing device 2, and provides a graphical user interface. The centralized management and control of the cloud management server breaks down the barriers of decentralized management of edge computing device 2. It can complete basic operations such as device registration and activation in batches, monitor the device's operating status in real time and distribute tasks remotely. At the same time, the graphical user interface reduces the complexity of management operations and improves the overall system management efficiency and convenience. The cloud management server includes a visual workflow editing module. This module allows users to define video analysis processing logic to generate pipeline configurations by dragging and dropping components in a graphical interface. The pipeline configurations are then remotely distributed to edge computing device 2 for execution. This visual operation method, which allows users to quickly define video analysis processing logic without needing complex coding skills, shortens the pipeline configuration design cycle. At the same time, the remote configuration distribution method enables unified updates and deployment of analysis logic on edge computing device 2, improving the flexibility and efficiency of system configuration.

[0019] Furthermore, the edge computing device 2 includes: The AI ​​inference module is used to load and execute deep learning models for object detection. The dedicated function design of the AI ​​inference module is specifically adapted to the loading and execution requirements of deep learning models, optimizes the allocation of computing resources for model inference, improves the speed and accuracy of object detection, and ensures the effectiveness and reliability of real-time analysis at the edge. The pipeline management module is used to receive, load, and execute pipeline configurations issued by the cloud management server to schedule the video acquisition, AI inference, and result processing processes. The pipeline management module can accurately parse the configuration instructions issued by the cloud and coordinate the execution order of video acquisition, AI inference, and result processing in an orderly manner. Communication module 3 is used to establish a two-way communication connection with the cloud management server via the MQTT protocol to receive remote commands and report device heartbeats and structured event data. The MQTT protocol is lightweight and highly reliable, and is adapted to the different network environments of edge computing device 2.

[0020] Furthermore, the pipeline configuration generated by the visual workflow editing module includes input source components, AI processing components, and output components connected in sequence; the AI ​​processing component is associated with a deep learning model for object detection, and the deep learning model for object detection is remotely hot-updated through a cloud management server; the modular structure of the pipeline allows each link to be adjusted and replaced independently, improving the flexibility and scalability of the pipeline configuration.

[0021] Furthermore, cloud management servers also include: The device management module manages the list, details, and geographical location of edge computing devices 2, and provides a remote command issuance interface. Remote commands include device restart, log collection, video streaming control, and firmware upgrade. The device management module can centrally store and display various device information, allowing maintenance personnel to quickly grasp the device status. At the same time, the rich remote command interface supports device fault diagnosis, configuration optimization, firmware updates, and other operations, enabling device maintenance to be completed without on-site maintenance, thus reducing the manpower and time costs of system maintenance. The data analysis dashboard module receives structured event data reported by edge computing device 2 and displays it in chart form as device operation indicators, biological target identification statistical trends and spatial distribution. The chart-based display can transform complex structured data into clear visual information, making it easier for operation and maintenance personnel to quickly interpret the device operation status, biological target identification patterns and spatial distribution characteristics. This provides intuitive and accurate data support for biological monitoring decisions in complex scenarios, improving the efficiency of data analysis and the scientific nature of decision-making.

[0022] Furthermore, it also includes mobile terminal applications, which communicate with the cloud management server for: Receive and display real-time event notifications generated by edge computing device 2 and forwarded by cloud management server; the portability of mobile terminal applications allows users to receive critical event notifications anytime, anywhere. View the real-time status, geographical location, and historical operating data of edge computing device 2; Request and play live video streams or recorded historical videos from the camera associated with edge computing device 2.

[0023] Furthermore, the mobile terminal application establishes a connection with the cloud management server via the WebSocket protocol to receive real-time event notifications. The types of real-time event notifications include target detection events, target capture events, device failure events, and device status change events. The full-duplex communication feature of the WebSocket protocol enables real-time data interaction between the cloud and the mobile terminal, ensuring the timely delivery of various event notifications, reducing the notification delay caused by polling, and adapting to the transmission requirements of multiple notification types to meet the event perception needs in different scenarios.

[0024] Furthermore, the edge computing device 2 automatically detects and switches between at least two networks, including wired networks, wireless LANs, and cellular mobile networks, and automatically reconnects to the cloud management server after network failure is recovered. The multi-network automatic switching function can adapt to network fluctuations and failures in complex scenarios, avoiding the loss of connection between the device and the cloud due to a single network failure. At the same time, the automatic reconnection mechanism after network recovery ensures that the device can quickly restore communication with the cloud, ensuring the continuity and stability of system monitoring.

[0025] Furthermore, the firmware upgrade package distributed by the cloud management server adopts a blue-green deployment method, and edge computing device 2 automatically rolls back to the previous software version when the upgrade fails; this reduces the risk of firmware upgrades for edge computing device 2 and ensures the stability of system operation.

[0026] Furthermore, the specific biological target is rodents, enabling rodent infestation monitoring.

[0027] Furthermore, this includes the following steps: On the cloud management server, the configuration of a video analysis pipeline for a specific biological target can be defined through a visual workflow editing module; The pipeline configuration and related AI models are remotely distributed from the cloud management server to the edge computing device 2; Edge computing device 2 loads and runs the pipeline configuration to perform real-time AI analysis on the accessed local video stream and identify specific biological targets; Edge computing device 2 reports the identified structured event data to the cloud management server; The cloud management server aggregates and displays event data from edge computing device 2, and pushes key event notifications to users through mobile terminal applications.

[0028] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection, an electrical connection, or a connection that allows communication between them; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components, unless otherwise explicitly limited. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0029] Obviously, the embodiments described above are merely some embodiments of the present invention, not all embodiments. The accompanying drawings show preferred embodiments of the present invention, but do not limit the patent scope of the present invention. The present invention can be implemented in many different forms; rather, these embodiments are provided to provide a more thorough and complete understanding of the disclosure of the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing specific embodiments, or make equivalent substitutions for some of the technical features. Any equivalent structures made using the content of this specification and drawings, directly or indirectly applied to other related technical fields, are similarly within the patent protection scope of this invention.

Claims

1. A biometric identification system for complex scenarios, characterized in that, include: Multiple cameras (1) are deployed at the monitoring site; Edge computing device (2) is used to access video streams collected by multiple cameras (1) and run an AI visual analysis model locally to perform real-time analysis of the video streams, identify specific biological targets, and generate structured event data containing target information; The cloud management server communicates with the edge computing device (2) and is used to uniformly register, activate, monitor the status, remotely distribute tasks and aggregate data for the edge computing device (2), and provides a graphical user interface; The cloud management server includes a visual workflow editing module, which allows users to define the processing logic of video analysis by dragging and dropping components in a graphical manner to generate pipeline configurations and remotely send the pipeline configurations to edge computing devices (2) for execution.

2. The biometric system for complex scenarios according to claim 1, characterized in that, The edge computing device (2) includes: The AI ​​inference module is used to load and execute deep learning models for object detection. The pipeline management module is used to receive, load, and execute pipeline configurations issued by the cloud management server to schedule video acquisition, AI inference, and result processing. The communication module (3) is used to establish a two-way communication connection with the cloud management server through the MQTT protocol to receive remote instructions and report device heartbeats and structured event data.

3. The biometric system for complex scenarios according to claim 2, characterized in that, The pipeline configuration generated by the visualization workflow editing module includes an input source component, an AI processing component, and an output component connected in sequence; the AI ​​processing component is associated with a target detection deep learning model, and the target detection deep learning model is remotely hot-updated through a cloud management server.

4. The biometric system for complex scenarios according to claim 1, characterized in that, The cloud management server also includes: The device management module is used to manage the list, details and geographical location of edge computing devices (2), and provides a remote command issuance interface, the remote commands including device restart, log collection, video streaming control and firmware upgrade; The data analysis dashboard module is used to receive the structured event data reported by the edge computing device (2) and display the device operation indicators, biometric target recognition statistical trends and spatial distribution in the form of charts.

5. The biometric system for complex scenarios according to claim 1, characterized in that, It also includes a mobile terminal application, which communicates with the cloud management server and is used for: Receive and display real-time event notifications generated by the edge computing device (2) and forwarded by the cloud management server; View the real-time status, geographical location and historical operating data of the edge computing device (2); Request and play the live video stream or recorded historical video from the camera associated with the edge computing device (2).

6. The biometric system for complex scenarios according to claim 5, characterized in that, The mobile terminal application establishes a connection with the cloud management server via the WebSocket protocol to receive the real-time event notifications; The types of real-time event notifications include target detection events, target capture events, equipment failure events, and equipment status change events.

7. The biometric system for complex scenarios according to claim 1, characterized in that, The edge computing device (2) automatically detects and switches between at least two networks, including wired network, wireless local area network and cellular mobile network, and automatically reconnects to the cloud management server after the network failure is recovered.

8. The biometric system for complex scenarios according to claim 1, characterized in that, The firmware upgrade package issued by the cloud management server adopts the blue-green deployment method, and the edge computing device (2) automatically rolls back to the software version before the upgrade when the upgrade fails.

9. The biometric system for complex scenarios according to claim 1, characterized in that, The specific biological target is rodents, used for rodent infestation monitoring.

10. The biometric system for complex scenarios according to claim 1, characterized in that, Includes the following steps: On the cloud management server, the configuration of a video analysis pipeline for a specific biological target can be defined through a visual workflow editing module; The pipeline configuration and related AI model are remotely downloaded from the cloud management server to the edge computing device (2). The edge computing device (2) loads and runs the pipeline configuration to perform real-time AI analysis on the accessed local video stream and identify specific biological targets; The edge computing device (2) reports the identified structured event data to the cloud management server; The cloud management server aggregates and displays event data from edge computing devices (2) and pushes key event notifications to users through mobile terminal applications.

Images