Automatic object tracking for an unmanned underwater vehicle using real-time image filtering and correlation

The focus of this research was automatic visual tracking of realistic underwater objects. Powerful real-time filtering and correlation hardware was used to generate full-image range and velocity maps at frame rates. These maps were in turn processed to provide object centroid position and rate. The object tracking system was tested experimentally on an underwater robotic vehicle, and qualitative results were obtained. Using this tracking system, the vehicle successfully tracked several different objects automatically. The feasibility of tracking realistic underwater objects in real-time was proven