There is a critical need to improve the situational awareness of the vulnerable parts of vehicles. Specifically, JIDA wants to deploy a system that uses vision based sensors to identify when a small object (minimum 1” cube) has been attached to the rear areas of a Toyota Land Cruiser’s side panels, and to the back panel. “Attached to a panel” means that a newly placed item is present for at least 1 second.

Our approach to detecting these anomalies on the surface of the vehicle is to use machine vision that is continuously monitoring the surface from enough vantage points to cover all of the vulnerable areas. The computational power necessary to perform detection using video recognition will be supported within a small form-factor PC. An in vehicle tablet will be used display the sensor feeds and generate user alerts. We are working directly the prime and the government to also support the integration efforts, testing, evaluation, and algorithm improvement.