AI-Integrated Hazardous Chemical Substance Detection System

Overview

The rapid increase in chemical products due to technological advancements in precision organic chemistry and semiconductors within the country has led to a continuous occurrence of major chemical accidents. Accidents involving hazardous chemicals have faced limitations in rapid response, relying on the experience or communication of on-site personnel, and proving inefficient in emergency situations. To enable swift response to hazardous chemical accidents and effectively reduce emergency response times, the development of an AI-based hazardous chemical identification system was pursued.

Project description

The National Fire Agency processed and trained data on 23 types of hazardous chemicals it possesses, developing an AI model based on environmental analysis that can automatically detect various elements—such as flames, smoke, solutions, and flashes—during fire incidents involving hazardous chemicals. An AI-integrated hazardous chemical detection system was established by adding certain features for field deployment, and this system was demonstrated at National Fire Agency client sites (including the Gyeongnam Fire Headquarters).

This AI solution automatically identifies hazardous chemicals during chemical disasters and includes the capability to visualize* the potential damage spread range. It provides a service that instantly shares the hazardous chemical identification results and response methods with the 119 Emergency Operations Center and field responders via the operations center screen and a mobile app.

*Firefighting response strategies (water dousing, spray extinguishing, suppression of combustible vapors after fire containment, etc.), prediction of damage scope including fire spread, securing early evacuation distances, and support for protective action decisions

Key technological features

• Developed an AI model based on analysis of the detection environment by training it on the types of hazardous chemicals and related data
• Captures disaster sites in real-time and identifies hazardous chemical types through image analysis technology
• Real-time delivery of information to fire scenes, including hazardous chemical substance details and response methods during accidents

Expected outcomes

The AI-integrated hazardous chemical detection system achieved a 94% recognition rate after three years of development and refinement. The AI solution was pilot-linked to the target system at the Gyeongnam Fire Headquarters, a demand site of the National Fire Agency.

The AI system was applied to the field. According to the results of the AI system's field testing and application, the National Fire Agency's hazardous chemical identification time decreased by 80% from 5 minutes to 1 minute, and the accuracy of hazardous chemical identification significantly improved. Furthermore, by delivering hazardous chemical information and response methods within one minute, it enables rapid incident response at fire scenes during hazardous chemical accidents. It is expected to enhance the safety of field personnel and ensure public safety by facilitating the initial evacuation of responding units and securing protective distances.

Launch year: 2023

Attribution

Case study courtesy of the National Information Society Agency, Republic of Korea, Apolitical's Content Partner. Sourced from Use Cases of Public AI Service, Vol.1,  https://eng.nia.or.kr/site/nia_eng/ex/bbs/View.do?cbIdx=31975&bcIdx=28997&parentSeq=29035