AI Detects Suicide Attempts at Hangang Bridges... Surveillance Technology Developed, Pilot Application from October

Seoul Institute of Technology and Seoul Fire & Disaster Headquarters Collaborate on 'Customized CCTV Surveillance Technology for Hangang Bridges'
Advancing Selective Surveillance for Detecting and Predicting Jumping Incidents on Hangang Bridges through AI-Based CCTV Video Learning

[Asia Economy Reporter Lim Cheol-young] The Seoul Institute of Technology announced on the 13th that it has completed the development of a data science-based ‘Han River Bridge Customized CCTV Surveillance Technology’ in collaboration with the Seoul Fire and Disaster Headquarters, and plans to pilot it starting in October.

According to the Seoul Institute of Technology, the institute will integrate the newly developed surveillance technology with the existing surveillance system and link it to the ‘Han River Bridge Integrated Control Center,’ scheduled to be established in October this year, for pilot application. After conducting demonstration tests until the end of the year, full-scale implementation is planned. As data accumulates, the accuracy of AI-based selective surveillance is expected to continuously improve.

The ‘Han River Bridge Integrated Control Center’ supports integrated command and control and systematic rescue activities when a jumping incident occurs on the Han River bridges. It is being established within the Ttukseom Water Rescue Team. Currently, individual surveillance is conducted by each water rescue team.

The newly developed surveillance technology uses artificial intelligence (AI) to learn the patterns of jump attemptors through video deep learning to detect and predict dangerous situations. Then, it ‘selectively’ displays only the CCTV footage of the relevant location on the surveillance operator’s monitor. Through AI-based selective surveillance, rapid response to situations before and after a jump is possible, minimizing monitoring gaps and improving surveillance efficiency.

The Seoul Institute of Technology expects that once the ‘Integrated Control Center’ is established and operated under the AI selective surveillance system, it will enhance proactive response capabilities to identify abnormal signs of jumping incidents and prevent jump attempts themselves. Water rescue team members will have reduced CCTV monitoring burdens and can focus more on rescue activities. Flexible responses will also be possible when multiple incidents occur simultaneously.

Ko In-seok, Director of the Seoul Institute of Technology, said, “This research is a collaborative case led by both institutions to minimize human casualties caused by jumping incidents on Han River bridges,” and added, “We will continue to conduct practical research using data science analysis technology for the safety and happiness of citizens.”