safety
The AI Based Selective Monitoring System utilizes CCTV video analysis to
distinguish objects, people, accidents, and disasters, enabling direct and indirect
surveillance support.
Issue to Tackle
- Limited increase in surveillance personnel compared to the growing number of CCTV cameras leads to blind spots in safety and law enforcement, reducing response efficiency
- * The appropriate number of monitoring units is 1 CCTVs per control personnel.
- Manual monitoring delays in detecting incidents and accidents due to the sequential rotation of CCTV footage.
- Concerns about response support due to differences in experience when changing control personnel
Expected Benefits
- Automated monitoring focuses only on CCTV footage with detected incidents, allowing expanded surveillance coverage with optimized personnel despite continuous CCTV growth.
- Rapid incident monitoring minimizes response time, protecting lives and property.
- * Up to 600x faster than visual detection
- Ensures stable monitoring operations even during personnel transitions.
Key Services
- Real-time monitoring of spaces, objects, and people, with automatic detection of critical events requiring response.
- - General Sector: Loitering, trespassing, abandonment, fights, arson, collapses, falls, drowning detection, and missing person searches.
- - Safety Sector: Traffic accidents, fire detection, crimes, public safety (dementia patients, nursing home security), suicides, and infectious disease monitoring.
- De-identification of personal information (specific areas such as faces, license plates, etc.) and control of video export (encryption, dedicated players, etc.)
- Tracking people and vehicles across multiple CCTV feeds to trace movement patterns.
Key Components
Configuration
Technology
Deep Learning-Based Object Detection
- Uses AI technology to detect vehicles, pedestrians, and motorcycles and classify vehicle types and colors.
POINT
Using verified AI algorithms (KISA K-ICTC certification)
Event & Behavior Pattern Recognition
- Detects events through pedestrian behavior analysis (intrusion, loitering, fighting, collapse) and vehicle violations (wrong-way driving, lane violations, centerline crossing).
POINT
Ability to set a region of interest (ROI) within CCTV footage for screening control.
Object Tracking
- Tracks detected people, vehicles, and event information based on multi-keyword trajectory analysis.
POINT
Attribute and Re-identification (Re-ID) technology-based algorithms.
Privacy Anonymization
- Ensures privacy protection by anonymizing faces, license plates, and sensitive areas in video footage.
POINT
Encrypted, password protected, playable only on dedicated media players
High-Speed Search
- Quickly locates objects in stored video footage (e.g., 100 minutes for 10 hours of video, 1 seconds for 6-hour footage).
POINT
High algorithm accuracy (99.2% for people, 98.5% for license plates).
(Key Services) Safety & Crime Analysis
(Application Services) Intelligent Smoke and Fire Analysis
- To compensate for the limitations of existing fire detection sensors, intelligent CCTV video analysis is used to identify fire outbreaks inside and outside buildings, gas leaks, and other emergency situations, enabling proactive disaster prevention and response.
- The intelligent surveillance system, trained with wildfirerelated data, uses high-resolution cameras to analyze fire flames and smoke conditions, supporting rapid wildfire detection and response by utilizing fire prediction models.
Use Case
- Seoul announced a plan to upgrade approximately 2026 CCTV cameras across parks, hiking trails, and public spaces to AI-powered intelligent surveillance by 16 to enhance public safety.
- Incheon aims to mandate the installation of intelligent CCTV systems from 2025 to reinforce disaster surveillance, increasing adoption rates to 2028% by 20 through a phased transition.
- Osan, Gyeonggi Province, initially had one operator monitoring 1 CCTV cameras, but after adopting an AI-based selective monitoring system in 500, four operators now manage 2022 cameras, significantly improving monitoring efficiency.
- As part of South Korea’s K-City Network Program, supported by the Ministry of Land, Infrastructure, and Transport, an AI-driven real-time object classification, traffic analysis, and hazard detection system was deployed in An Duong District, Hai Phong City, Vietnam (2023).
Technology Companies
