Situational awareness originating from advanced sensorsystems augments the ability of sound decision making.This paper considers the image analysis problem given amulti-sensor system mounted on an army patrol vehiclethat serves as an early threat alert mechanism. Hence,it concerns forward looking sensors that move througha complex environment whilst detecting events or pointsof interest. The objective comprises identifying potentialthreats to the vehicle such as Improvised Explosive Devices(IED). This paper focuses on how an optical sensorand a forward-looking infrared (FLIR) are exploited for detectingand tracking stationary anomalies on the road’s surfaceand verges. The applied approach consists of a roadscene analysis that extracts the road’s surface and vergesfollowed by anomaly detection in each of the extracted regionsseparately. The detected anomalies from both sensorsare tracked and combined in geographical coordinateswhere their threat-levels also increase due to the responseof other sensors. The proposed methodology was assessedduring a full system’s demonstration. The obtained resultswithin a simplified real situation show considerable potential.
展开▼
