Korea South-East Power
VLM-based CCTV safety monitoring
Korea South-East Power reported at least 90% detection accuracy from AI safety cameras
Problem
Power-plant safety monitoring concentrated on selected high-risk work areas. Staff needed earlier warnings of dangerous situations.
Approach
Korea South-East Power introduced CCTV analysis using a vision-language model. The system processed images and text to identify abnormal situations and issue alerts. The company reported results from the deployed system while planning a wider camera rollout.
Results
Takeaway
Detection accuracy described the monitoring system's performance. It did not establish a reduction in workplace accidents.
Categories
- Size
- Large enterprise
- Industry
- Energy and utilities
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Computer visionGenerative AI
- Adoption
- Vendor solution
- Goal
- Quality and risk
- Evidence
- Company's own claim
Sources
Source summaries
Korea South-East Power reported detection accuracy of at least 90% after introducing VLM-based video analysis. Expansion across its full CCTV network remained a plan at the announcement.
Full details are in the original sources.