Intelligent security early warning method and device based on TOF posture recognition

Abstract

The invention relates to the field of intelligent security and protection, in particular to an intelligent security and protection early warning method and device based on TOF posture recognition, and the device comprises a monitoring camera instrument with a TOF sensing technology, a cloud server which is used for timely sending out alarm information by recognizing the posture of a human body, a monitoring terminal which can receive alarm signals in real time, and the like. Compared with the prior art, the device has the advantages that firstly, the TOF infrared light pulse sensing imaging technology is used for shooting crowds in the detected environment, obtained pictures do not relate to the privacy of students, but whether fighting happens or not is judged by recognizing human body postures in combination with the HAR technology; after a campus bullying event occurs, related monitoring personnel can be warned and notified to arrive at the scene for stopping and processing through the scene; on the basis, the device and the technical scheme can also be used in various fields such as household drop resistance of the old, public privacy area security and protection, industrial safety production and the like.

Description

Technical Field

[0001] The present invention relates to the field of intelligent security, and in particular to an intelligent security early warning method and device based on TOF gesture recognition. Background Art

[0003] School bullying usually occurs in relatively hidden places, such as school toilets, stair corners with low traffic, and alleys outside the school. These are usually blind spots for surveillance. Moreover, even if bullying incidents are detected in time, no one will discover them in time to effectively stop them and impose appropriate punishment on the perpetrators, which makes the perpetrators feel lucky and become more aggressive. This is because due to relevant regulations, traditional optical cameras will infringe on students' privacy. It is relatively awkward to arrange surveillance cameras in all directions in places such as toilets, and it cannot be effectively implemented. Therefore, an intelligent security warning method and device based on TOF gesture recognition are provided. Summary of the invention

[0004] 1. Technical Problems Solved

[0005] The technical problem to be solved by the present invention is that when using the monitoring effect of cameras to reduce and prevent school bullying, traditional cameras are inconvenient to set up in some special places, and simply recording by shooting cannot effectively prevent the perpetrators and protect the victims.

[0006] 2. Technical Solution

[0007] In order to solve the above technical problems, the technical solution provided by the present invention is: an intelligent security warning method and device based on TOF gesture recognition, including a monitoring camera, which is used to shoot and monitor the inspected area.

[0008] The monitoring camera contains a TOF sensor for collecting human posture through infrared light pulse shooting. The monitoring camera is respectively equipped with a light transmitter and a light detector for transmitting and receiving infrared light pulse signals. By deploying a multi-node TOF sensor network, a millimeter-level precision depth point cloud is generated.

[0009] It also includes a cloud server and an information feedback terminal. The front end and back end of the cloud server are respectively interconnected with the monitoring camera and the information feedback terminal. The cloud server uses AI-based HAR technology to recognize human posture, analyzes multi-person interaction postures through a spatiotemporal graph neural network, extracts risk features, calculates the bullying risk index, and triggers a graded response.

[0010] Furthermore, the monitoring camera is installed on the wall or ceiling of the inspected area in a ceiling-mounted manner.

[0011] Furthermore, the surveillance camera lens containing TOF sensing technology has a resolution setting of ≥1024×768.

[0012] Furthermore, the cloud server includes two types: system shared cloud and institution private cloud.

[0013] Furthermore, the cloud server contains a data processor carrying an expert system and a shared knowledge base carrying a deep learning system. The shared knowledge base serves as a supplement to the recognition capability of the data processor, and constitutes an update model of the federated learning framework by combining the deep learning system and several expert system units.

[0014] Furthermore, a warning device cooperating with the cloud server is installed in the inspected area, and the warning device includes an alarm with sound and light alarm functions and a speaker with voice broadcast function.

[0015] Furthermore, the information feedback terminal includes a centralized system terminal platform interface and a personal mobile APP alarm interface.

[0016] 3. Beneficial Effects

[0017] The advantages of the present invention compared with the prior art are: firstly, the device uses TOF infrared light pulse sensing imaging technology to shoot the crowd in the inspected environment, and the obtained pictures will not involve the privacy of students, but combined with HAR technology, it can identify human posture to determine whether a fight has occurred. Secondly, after a school bullying incident occurs, the system of the device can issue on-site warnings and notify relevant monitoring personnel to arrive at the scene to stop and deal with it; based on this, the device and the technical solution can also be used in multiple fields such as home fall protection for the elderly, security in public privacy areas, and industrial safety production. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 The present invention is a flowchart of an intelligent security warning method and device based on TOF posture recognition during warning action.

[0019] Figure 2 It is a schematic diagram of the appearance structure of an intelligent security warning device based on TOF gesture recognition of the present invention.

[0020] Figure 3 It is a schematic diagram of an intelligent security warning method and device information collection method based on TOF posture recognition of the present invention.

[0021] Figure 4 The present invention is an operation flow chart of an intelligent security warning method and device based on TOF posture recognition when triggering campus anti-bullying monitoring warning.

[0022] As shown in the figure: 1. Monitoring camera, 2. Light generator, 3. Light detector. DETAILED DESCRIPTION

[0023] The present invention is further described in detail below in conjunction with the accompanying drawings.

[0024] In order to solve the above technical problems, the technical solution provided by the present invention is: Figure 2 , an intelligent security warning method and device based on TOF gesture recognition, including a monitoring camera 1, through which the monitoring camera 1 is used to shoot and monitor the inspected area, and the monitoring camera 1 is installed on the wall or ceiling of the inspected area in a ceiling-mounted manner;

[0025] The place to be filmed and monitored is called the inspected area. Installing a monitoring camera 1 in the place can effectively carry out safety prevention and facilitate the acquisition of on-site or historical information;

[0026] Combined with Figure 3 The monitoring camera 1 contains a TOF sensor for collecting human posture through infrared light pulse shooting. The monitoring camera lens containing TOF sensing technology has a resolution of ≥1024×768. The monitoring camera 1 is respectively equipped with a light transmitter and a light detector 3 for transmitting and receiving infrared light pulse signals. By deploying a multi-node TOF sensor network, a millimeter-level precision depth point cloud is generated;

[0027] The TOF sensor generates high-precision depth images by emitting infrared light pulses and measuring the return time of the light pulses. The TOF sensing imaging technology has the advantages of high precision, high privacy, and good anti-interference effect. It can capture accurate depth information and generate 3D images. It is not affected by environmental factors and can work stably for a long time under various conditions. It does not need to capture visible light images to ensure the privacy of the subject. It can also respond quickly when emergencies occur, and is suitable for real-time monitoring and alarm.

[0028] Combined with Figure 1 , and also includes a cloud server and an information feedback terminal. The cloud server includes two types: a system shared cloud and an institutional private cloud. The system shared cloud can be associated with multiple independent departments in a wide area, and connect to the monitoring cameras 1 in each inspected area in each department. The institutional private cloud belongs to a certain department and serves a specific department. It connects with the monitoring cameras 1 in each inspected area in the department, which is convenient for individuals to adapt conditions and meet private or special conditions and scene needs;

[0029] The front end and back end of the cloud server are respectively associated with the monitoring camera 1 and the information feedback terminal. The cloud server uses the AI-based HAR technology to recognize human postures, analyzes multi-person interactive postures through the spatiotemporal graph neural network, extracts risk features, calculates the bullying risk index, and triggers a graded response;

[0030] The cloud server contains a data processor with an expert system and a shared knowledge base with a deep learning system. The shared knowledge base is set as a supplement to the recognition capability of the data processor, and the deep learning system and several expert system units are combined to form an update model of the federated learning framework.

[0031] The comprehensive recognition effect of the intelligent algorithm combining expert system and deep learning is significantly better than that of a single model algorithm. The expert system algorithm can model, track motion, and analyze postures of space, objects, human limbs, and extremities. When the expert system cannot recognize the target object, the deep learning system intervenes.

[0032] The deep learning system uses the trained neural network model to well identify objects such as human head, limbs, and extremities, and pass the identified targets to the expert system to continue motion tracking and posture analysis;

[0033] The information feedback terminal includes a centralized system terminal platform interface and a personal mobile APP alarm interface, wherein the system terminal platform interface is provided by the relevant product supplier and further provides alarm information transmission services, while the mobile APP alarm interface corresponds to a method of notifying the alarm information to a specific person or terminal in a targeted manner;

[0034] The cloud server and information feedback terminal include four association modes, namely, system shared cloud-system terminal platform interface, system shared cloud-mobile APP alarm interface, organization private cloud-system terminal platform interface, organization private cloud-mobile APP alarm interface, and on the basis of the above association modes, one or more association modes are selected to jointly build a security early warning system;

[0035] A warning device that cooperates with the cloud server is also installed in the inspected area, and the warning device includes an alarm with sound and light alarm function and a sound with voice broadcast function;

[0036] In addition to rushing to the scene and notifying security and police after receiving manufacturer notifications or APP alarm information, in certain scenarios, such as school bullying, monitoring personnel can also promptly use warning devices installed on site to automatically broadcast warning signals to stop and deter perpetrators.

[0037] The monitoring camera 1 of the device has a recognition distance of 0.6-8 meters and a recognition area of ​​25-30 square meters. It uses a 3D laser radar and can identify human postures such as falling, fighting, gathering, and staying. Its cloud server can realize network, Bluetooth and other transmission functions through related equipment, and continuously and all-weather monitor the environmental status of the inspected area.

[0038] In the specific implementation of the present invention, when the device is installed and equipped in the school bathroom for campus bullying monitoring and early warning, the monitoring camera 1 is installed above the appropriate area of ​​the bathroom. When the monitoring camera TOF sensor radar detects signs of fighting, the school's supporting institution private cloud first uses the sound and light alarm and intelligent voice audio and other equipment installed in the bathroom to issue an alarm whistle and dissuasion and warning recordings. The monitoring camera 1 observes the posture and movement of the people in the inspected area in real time. If the violent behavior is fortunately prevented, the alarm device will pause, and the cloud server will record the captured information of the relevant personnel as a backup. If the perpetrator ignores the sound and light warning and continues to commit campus violence, the cloud server will generate a work order and send the relevant information of this behavior to the monitoring personnel's mobile phone APP. The monitoring personnel can be school security guards, teachers from the moral education department, etc. After receiving the alarm information, the relevant personnel rush to the scene to stop and deal with it. This avoids the continued violence of the victimized students, and the campus violence incident cannot be exposed and handled. This process can be combined with the attached Figure 4 For reading, since the monitoring camera 1 adopts infrared light pulse shooting and information collection, to a certain extent, the information of relevant personnel involved in campus violence incidents can also be effectively protected to avoid secondary harm.

[0039] The present invention and its embodiments are described above, and such description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design a structure and embodiment similar to the technical solution without creativity, which should belong to the protection scope of the present invention.

Claims

1. An intelligent security warning method and device based on TOF gesture recognition, comprising a monitoring camera (1), wherein the monitoring camera (1) is used to photograph and monitor the inspected area, and is characterized in that: The monitoring camera (1) contains a TOF sensor for collecting human posture through infrared light pulse shooting. The monitoring camera (1) is respectively equipped with a light transmitter and a light detector (3) for transmitting and receiving infrared light pulse signals. A multi-node TOF sensor network is deployed to generate a millimeter-level precision depth point cloud. It also includes a cloud server and an information feedback terminal. The front end and back end of the cloud server are respectively connected to the monitoring camera (1) and the information feedback terminal. The cloud server uses the AI-based HAR technology to recognize human body postures, analyzes multi-person interactive postures through a spatiotemporal graph neural network, extracts risk features, calculates the bullying risk index, and triggers a graded response.

2. According to claim 1, an intelligent security warning method and device based on TOF gesture recognition is characterized in that: The monitoring camera (1) is installed on the wall or ceiling of the inspected area in a ceiling-mounted manner.

3. According to claim 2, an intelligent security warning method and device based on TOF gesture recognition is characterized in that: The monitoring camera lens containing TOF sensing technology has a resolution setting of ≥1024×768.

4. According to claim 1, an intelligent security warning method and device based on TOF gesture recognition is characterized in that: The cloud servers include two types: system shared cloud and institution private cloud.

5. According to claim 4, an intelligent security warning method and device based on TOF gesture recognition is characterized in that: The cloud server contains a data processor carrying an expert system and a shared knowledge base carrying a deep learning system. The shared knowledge base serves as a supplement to the recognition capability of the data processor, and constitutes an update model of the federated learning framework by combining the deep learning system and several expert system units.

6. According to claim 5, an intelligent security warning method and device based on TOF gesture recognition is characterized in that: A warning device that cooperates with the cloud server is also installed in the inspected area, and the warning device includes an alarm with sound and light alarm functions and a speaker with voice broadcast function.

7. According to claim 1, an intelligent security warning method and device based on TOF gesture recognition is characterized in that: The information feedback terminal includes a centralized system terminal platform interface and a personal mobile APP alarm interface.

Images