An Integrated Fusion Engine for Early Threat Detection Demonstrated in Public-Space Trials

Area of Science:

• Security engineering
• Computer vision
• Data fusion

Background:

• Counter-terrorism in public spaces presents significant challenges.
• Early detection of threats such as weapons or explosives is crucial for public safety.
• Existing security measures can disrupt the flow of commuters in high-traffic areas.

Purpose of the Study:

• To develop and evaluate an integrated sensor fusion engine for real-time threat detection in public spaces.
• To assess the system's effectiveness in improving security for soft targets without impeding commuter flow.
• To validate the system's performance through a public-space trial.

Main Methods:

• Development of an integrated fusion engine managing multiple sensors for threat detection.
• Implementation of two data pipelines: one for smart glasses visualization and another for sensor hand-over.
• Integration of 2D-video tracking, person re-identification, 3D-video tracking, and command and control (C&C) components.
• Conducting a trial in a Rome metro station involving 586 commuters.

Main Results:

• The integrated fusion engine demonstrated high overall accuracy: 97.4% for the visualization pipeline and 97.6% for the hand-over pipeline.
• Individual system components achieved excellent accuracy: 2D Video (98.0%), Person Re-identification (100.0%), 3D Video (99.7%), and Command and Control (99.5%).
• The system successfully provided real-time threat information, including image and position data, to security operators.

Conclusions:

• The developed sensor fusion engine is highly effective for enhancing security in public spaces.
• The system can detect threats with high accuracy without negatively impacting commuter movement.
• This technology offers a significant improvement for securing soft targets like airports and railway stations.