Mission Flexible AI-Powered Video Analytics For Base Security

Threats to Air Force bases are growing more frequent, sophisticated, and deadly. Video Analytics and AI can leverage video and sensor data to address threats. Relying on busy operators to monitor video has created ‘information overload’. Our Video Processing and Modeling Tool (VPMT) continuously monitors and evaluates video, alerting operators to the few, critical issues requiring attention. VPMT is designed for security teams, tailorable to different use cases, locations, and infrastructure. We propose extending VPMT – advancing existing technologies -- to address the USAF Force Protection mission. Prior VPMT use includes large commercial venues, transit stations, and border security, with use cases such as detecting abandoned objects, tracking vehicles and associated people, and identifying unauthorized access. Other tools focus on narrow use cases and are not easily tailored. VPMT’s modular, flexible framework can quickly adapt models for new USAF uses. We identify the questions for useful security help, determine how to leverage current sensor data, and highlight critical issues for immediate or after-action investigation. Design is informed by data, research, and SME input. VPMT works with existing infrastructure. We use proven, calibrated algorithms with common languages, libraries, and tools. We iteratively develop, test, and validate new models for each use case.