CN-121981868-A - Method, device, medium and equipment for disaster response decision based on multiple modes

CN121981868ACN 121981868 ACN121981868 ACN 121981868ACN-121981868-A

Abstract

The application relates to the technical field of data processing, and particularly provides a method, a device, a medium and equipment for disaster response decision based on multiple modes, wherein the method can comprise the steps of acquiring structured data corresponding to original data of a disaster site in real time; the method comprises the steps of obtaining original data, namely video data, image data, voice data or disaster monitoring data, obtaining structured data, namely disaster event type, spatial position, confidence and time stamp, locating the semantic position of a disaster scene from a disaster chain knowledge base constructed in advance by taking the structured data as a retrieval element, reasoning a chain evolution path of the disaster scene, obtaining a response measure set corresponding to the semantic position, and obtaining a disaster response strategy according to the original data, the structured data, the chain evolution path and the response measure set. The embodiment of the application can improve the decision accuracy of disaster treatment.

Inventors

RUI JING
ZHENG JIANCHUN
WANG YUNTAO
YOU QIUJU
ZHONG SHAOBO
Shang Qiujin

Assignees

北京市科学技术研究院

Dates

Publication Date: 20260505
Application Date: 20260205

Claims (10)

1. A method of disaster response decision based on multiple modes, comprising: the method comprises the steps of acquiring structured data corresponding to original data of a disaster scene in real time, wherein the original data comprise video data, image data, voice data or disaster monitoring data, and the structured data comprise disaster event types, spatial positions, confidence and time stamps; The structural data is taken as a retrieval element, the semantic position of the disaster site is positioned from a pre-constructed disaster chain knowledge base, and a chain evolution path of the disaster site is inferred; acquiring a response measure set corresponding to the semantic position; And acquiring a disaster response strategy according to the original data, the structured data, the chained evolution path and the response measure set.
2. The method of claim 1, wherein the acquiring, in real time, the structured data corresponding to the raw data of the disaster site comprises: The method comprises the steps of obtaining structured data sent by multiple classes of edge equipment, wherein a multi-mode model is deployed on each class of edge equipment in the multiple classes of edge equipment, and the multi-mode model is used for carrying out structural conversion on event characteristics obtained by analyzing the original data collected by each class of edge equipment in real time to obtain the structured data.
3. The method of claim 1, wherein the acquiring, in real time, the structured data corresponding to the raw data of the disaster site comprises: Acquiring the original data after preprocessing the acquired initial data by multiple types of edge equipment; analyzing the original data in real time by using a multi-mode model to obtain event characteristics; and carrying out structuring treatment on the event characteristics to obtain the structured data.
4. A method according to any one of claims 2-3, wherein said reasoning about a chain-like evolution path of the disaster site comprises: Acquiring historical disaster data corresponding to the disaster site; And analyzing the historical disaster data and the event characteristics to obtain the chain evolution path.
5. A method according to any of claims 1-3, wherein said obtaining a disaster response strategy from said raw data, said structured data, said chained evolution path and said set of response measures comprises: inputting real-time text or abstract information corresponding to the original data into a first model to obtain a first response strategy; Inputting the structured data, the chained evolution path and the response measures into a second model to obtain a second response strategy; and determining the disaster response strategy through the first response strategy and the second response strategy, wherein the disaster response strategy is used for providing reference for field experts.
6. A method according to any of claims 1-3, wherein said obtaining a set of response measures corresponding to said semantic locations comprises: querying the response measure set corresponding to the semantic position from a pre-constructed disaster response tree.
7. The method of any of claims 1-3, wherein the disaster chain knowledge base is constructed by joint analysis of historical multimodal data comprising text data, picture data, and video data of historical disasters.
8. An apparatus for multimodal-based disaster response decision comprising: The system comprises a real-time acquisition module, a real-time acquisition module and a real-time processing module, wherein the real-time acquisition module is used for acquiring structured data corresponding to original data of a disaster scene in real time, wherein the original data comprises video data, image data, voice data or disaster monitoring data; The positioning reasoning module is used for positioning the semantic position of the disaster site from a pre-constructed disaster chain knowledge base by taking the structured data as a retrieval element and reasoning a chain evolution path of the disaster site; The measure acquisition module is used for acquiring a response measure set corresponding to the semantic position; And the decision module is used for acquiring a disaster response strategy according to the original data, the structured data, the chained evolution path and the response measure set.
9. A computer readable storage medium, characterized in that the computer readable storage medium has stored thereon a computer program, wherein the computer program when run by a processor performs the method according to any of claims 1-7.
10. An electronic device comprising a memory, a processor, and a computer program stored on the memory and running on the processor, wherein the computer program when run by the processor performs the method of any one of claims 1-7.

Description

Method, device, medium and equipment for disaster response decision based on multiple modes Technical Field The application relates to the technical field of data processing, in particular to a method, a device, a medium and equipment for disaster response decision based on multiple modes. Background Natural and man-made disasters can lead to serious casualties and economic losses, and thus timely effective disaster response is critical to reducing losses. Currently, when facing disasters, traditional disaster response decisions mainly depend on the decision experience of experts on historical disasters and judgment on site conditions. With the continuous development of artificial intelligence, analysis of disaster area sites can be realized by training a model through historical data, and then decision suggestions are given by experts. However, due to limitations of data sources and uncertainty of field conditions, the prior art cannot realize real-time monitoring of disaster areas on site, and cannot formulate an accurate response strategy. Therefore, how to provide a more accurate method for disaster response decision based on multiple modes is a technical problem to be solved. Disclosure of Invention The technical scheme of the embodiment of the application can comprehensively formulate a disaster response strategy with higher precision in real time by multi-mode data, realizes real-time response and treatment of disasters, and effectively avoids safety problems caused by secondary disasters. According to the method, structural data corresponding to original data of a disaster site are obtained in real time, the original data comprise video data, image data, voice data or disaster monitoring data, the structural data comprise disaster event types, spatial positions, confidence and time stamps, the structural data are used as search elements, semantic positions of the disaster site are located from a pre-built disaster chain knowledge base, a chain evolution path of the disaster site is inferred, a response measure set corresponding to the semantic positions is obtained, and a disaster response strategy is obtained according to the original data, the structural data, the chain evolution path and the response measure set. According to some embodiments of the application, after structured data corresponding to original data of a disaster scene are obtained in real time, semantic positions are positioned from a disaster chain knowledge base, chain evolution paths are inferred, then a response measure set is obtained through the semantic positions, and finally a disaster response strategy is obtained through the original data, the structured data, the chain evolution paths and the response measure set. The embodiment of the application can comprehensively formulate the disaster response strategy with higher precision by real-time multi-mode data, realize real-time response and treatment of disasters, and effectively avoid the safety problem caused by secondary disasters. In some embodiments, the real-time acquisition of the structured data corresponding to the original data of the disaster scene comprises the steps of acquiring the structured data sent by multiple types of edge devices, wherein a multi-mode model is deployed on each type of edge device in the multiple types of edge devices, and the multi-mode model is used for carrying out structural conversion on event characteristics obtained by carrying out real-time analysis on the original data acquired by each type of edge device to obtain the structured data. According to some embodiments of the application, the multi-mode model is used for processing the data acquired by the edge devices of various types to obtain the structured data, so that effective basic data is provided for the establishment of a subsequent disaster response strategy. In some embodiments, the real-time acquisition of the structured data corresponding to the original data of the disaster scene comprises the steps of acquiring the original data after preprocessing the acquired original data by the multi-class edge equipment, analyzing the original data in real time by using a multi-mode model to obtain event characteristics, and carrying out structuring processing on the event characteristics to obtain the structured data. According to the method, the device and the system, the original data sent by the multi-type edge devices are received, the multi-mode model is used for analyzing and structuring the original data to obtain structured data, and effective basic data is provided for the establishment of a follow-up disaster response strategy. In some embodiments, the reasoning the chained evolution path of the disaster site comprises obtaining historical disaster data corresponding to the disaster site, and analyzing the historical disaster data and the event characteristics to obtain the chained evolution path. According to some embodiments of the application, the chain type evol