OBJECT TRACKING USING MULTI-SOURCE PERSISTENT SURVEILLANCE
Richard L. Barrett, Christopher M. Camp, Nicholas W. Knize, Richard L. Landis · 2006
Tracking moving objects through remote sensing using Doppler shift is a well-established technique of radar technology. Once the moving objects that are of interest are identified, revisiting every few seconds allows tracking the objects’ paths. Similarly, video imagery can be used to track objects of interest using electro-optical change detection to identify and track moving objects. However, tracking an object by maintaining continuous surveillance of each object by a dedicated sensor is a very expensive use of sensor resources. Techniques are presented that support persistent surveillance by using multiple sources and types of remotely sensed data to track objects of interest. Longer gaps in remote sensor observations are possible using electro-optical change detection and by automatically determining revisit rates. Techniques are presented for improving geolocation and track fusion using pre-existing geopolitical information and object characterization, in addition to the current remotely sensed data. Temporal geolocation techniques support forensic analysis for law enforcement, traffic management, and military intelligence applications.