CN-121999573-A - Dangerous detection method and system

Abstract

The application relates to a hazard detection method and system. The method is applied to a first terminal and comprises the steps of receiving real-time audio data sent by a second terminal, wherein the real-time audio data are sent by the second terminal after monitoring preset keywords, the preset keywords are used for representing the current environment where the second terminal is located, detecting the current environment according to the real-time audio data, controlling the second terminal to collect real-time video data in response to the fact that the second terminal is detected to be dangerous, and sending dangerous detection data to a server, wherein the dangerous detection data at least comprise real-time video data, and the dangerous detection data are used for triggering the server to detect the current environment. By adopting the method, the danger detection can be automatically triggered, the real-time recognition of the danger is realized, and the effective protection can be obtained when the user cannot operate autonomously.

Inventors

• Xiang Guanjie
• HAN LIANG

Assignees

• 展讯通信(天津)有限公司

Dates

Publication Date

20260508

Application Date

20260311

Claims (10)

1. 1. A hazard detection method, applied to a first terminal, comprising: receiving real-time audio data sent by a second terminal, wherein the real-time audio data are sent by the second terminal after monitoring preset keywords, and the preset keywords are used for representing that the current environment where the second terminal is located has danger; detecting the current environment according to the real-time audio data; Controlling the second terminal to perform real-time video data acquisition in response to detecting the existence of the hazard, and And sending hazard detection data to a server, wherein the hazard detection data at least comprises the real-time video data, and the hazard detection data is used for triggering the server to detect the current environment.
2. 2. The method of claim 1, wherein detecting the current environment from the real-time audio data comprises: According to the real-time audio data, emotion detection and event detection are carried out, and a first detection result is obtained; and responding to the first detection result to represent that a preset emotion exists and/or a preset dangerous event exists, and determining that the current environment is dangerous.
3. 3. The method of claim 1, wherein the sending hazard detection data to the server comprises: Acquiring real-time motion data acquired by a motion sensor integrated in the first terminal; and sending hazard detection data comprising the real-time motion data and the real-time video data to a server.
4. 4. A method according to any one of claims 1-3, wherein the method further comprises: The method comprises the steps of acquiring real-time positioning information acquired by a positioning sensor integrated in the first terminal, and sending the real-time positioning information to the server, wherein the real-time positioning information is used for enabling the server to send alarm information comprising the real-time positioning information to a preset receiver in response to the fact that danger is detected.
5. 5. A hazard detection method, applied to a server, comprising: receiving dangerous detection data sent by a first terminal, and extracting visual characteristics of real-time video data in the dangerous detection data; Performing target detection on the current environment where the second terminal is positioned according to the visual characteristics to obtain a second detection result, wherein the target detection comprises at least one of scene detection, tool detection and attack behavior detection; Determining the dangerous existence condition of the current environment according to the second detection result; The real-time video data are collected by the first terminal after detecting that the current environment is dangerous according to real-time audio data, the real-time audio data are sent to the first terminal after the second terminal monitors preset keywords, and the preset keywords are used for representing that the current environment where the second terminal is located is dangerous.
6. 6. The method of claim 5, wherein the second detection result comprises a scene detection result, a tool detection result, and a behavior detection result, wherein the hazard detection data further comprises real-time motion data collected by a positioning sensor integrated in the first terminal, and wherein the method further comprises: According to the real-time motion data, carrying out gesture detection on the first terminal to obtain a gesture detection result; correspondingly, the determining, according to the second detection result, the dangerous existence condition of the current environment includes: weighting the gesture detection result, the scene detection result, the tool detection result and the behavior detection result according to preset weights respectively corresponding to gesture detection, scene detection, tool detection and attack behavior detection to obtain an environmental risk score; And determining the dangerous existence condition of the current environment according to the environmental risk score and the preset score threshold corresponding to each risk level.
7. 7. The method according to claim 5 or 6, characterized in that the method further comprises: receiving real-time positioning information sent by the first terminal, wherein the real-time positioning information is acquired by a positioning sensor integrated in the first terminal; The method further comprises the steps of: And sending alarm information to a preset receiver in response to the danger existing in the current environment, wherein the alarm information at least comprises the real-time positioning information and the danger information, and the danger information is determined based on the real-time video data.
8. 8. A hazard detection method, applied to a second terminal, comprising: monitoring the current environment where the second terminal is located; in response to monitoring effective audio data, detecting preset keywords of the effective audio data; Responding to the effective audio data comprising at least one preset keyword, collecting real-time audio data and sending the real-time audio data to a first terminal, wherein the preset keyword is used for representing that the current environment where a second terminal is located has danger; Responding to the control of the first terminal after detecting that the current environment has danger based on the real-time audio data, acquiring real-time video data, and sending the real-time video data to the first terminal; the real-time video data are used for triggering the first terminal to send hazard detection data to a server, the hazard detection data at least comprise the real-time video data, and the hazard detection data are used for triggering the server to detect the current environment.
9. 9. The method of claim 8, wherein the performing preset keyword detection on the valid audio data in response to listening to the valid audio data comprises: In response to monitoring effective audio data, inputting the effective audio data into a first audio detection model to obtain a third detection result; responding to the third detection result to represent that at least one preset keyword is detected, and inputting the effective audio data into a second audio detection model to obtain a fourth detection result, wherein the model structures of the first audio detection model and the second audio detection model are different; And responding to the fourth detection result representation to detect at least one preset keyword, and determining that the effective audio data comprises the at least one preset keyword.
10. 10. A hazard detection system, characterized in that the system comprises a first terminal for performing the method of any of claims 1-4, a second terminal for performing the method of any of claims 5-7, and a server for performing the method of any of claims 8-9.

Description

Dangerous detection method and system Technical Field The application relates to the technical field of safety protection, in particular to a hazard detection method and system. Background Personal safety protection demands are increasing, especially in the scenes of night travel, solitary, emergency medical treatment and the like. Portable wearable devices are widely used for security monitoring and emergency alerting due to their wearability and first view acquisition capability. In the traditional technology, when the portable wearable device is adopted to realize safety protection, dangerous detection is triggered mainly through a manual triggering (such as a button, a gesture and the like) mode. However, the conventional triggering mode has obvious bottleneck that the manual triggering cannot be operated when a user is stressed, injured or suddenly ill, so that dangerous signals cannot be sent out in time, and rescue cannot be achieved quickly. Disclosure of Invention Based on the above, it is necessary to provide a hazard detection method and system, which can automatically trigger hazard detection, realize real-time hazard identification, and ensure that a user can still obtain effective protection when the user cannot operate autonomously. In a first aspect, the present application provides a hazard detection method, applied to a first terminal, the method comprising: the method comprises the steps of receiving real-time audio data sent by a second terminal, wherein the real-time audio data are sent by the second terminal after monitoring preset keywords, and the preset keywords are used for representing that the current environment where the second terminal is located has danger; detecting a current environment according to the real-time audio data; controlling the second terminal to perform real-time video data acquisition in response to detecting the presence of a hazard, and And sending hazard detection data to the server, wherein the hazard detection data at least comprises real-time video data, and the hazard detection data is used for triggering the server to detect the current environment. In one embodiment, the current environment is detected according to the real-time audio data, and the method comprises the steps of carrying out emotion detection and event detection according to the real-time audio data to obtain a first detection result, and determining that the current environment is dangerous according to the fact that the first detection result represents that a preset emotion exists and/or a preset dangerous event exists. In one embodiment, the sending of the hazard detection data to the server includes obtaining real-time motion data collected by a motion sensor integrated in the first terminal and sending hazard detection data including real-time motion data and real-time video data to the server. In one embodiment, the method further comprises the steps of obtaining real-time positioning information acquired by a positioning sensor integrated in the first terminal and sending the real-time positioning information to the server, wherein the real-time positioning information is used for enabling the server to send alarm information comprising the real-time positioning information to a preset receiver in response to the fact that the danger is detected. In a second aspect, the present application also provides a risk detection method, applied to a server, where the method includes: Receiving dangerous detection data sent by a first terminal, and extracting visual characteristics of real-time video data in the dangerous detection data; performing target detection on the current environment where the second terminal is positioned according to the visual characteristics to obtain a second detection result, wherein the target detection comprises at least one of scene detection, tool detection and attack behavior detection; Determining the dangerous existence condition of the current environment according to the second detection result; The real-time video data are collected by the first terminal after detecting that the current environment is dangerous according to the real-time audio data, the real-time audio data are sent to the first terminal after the second terminal monitors preset keywords, and the preset keywords are used for representing that the current environment where the second terminal is located is dangerous. In one embodiment, the second detection result comprises a scene detection result, a tool detection result and a behavior detection result, the danger detection data further comprises real-time motion data collected by a positioning sensor integrated in the first terminal, the method further comprises the steps of carrying out gesture detection on the first terminal according to the real-time motion data to obtain a gesture detection result, and correspondingly, determining the dangerous existence of the current environment according to the second detection result, wh