Vision aided navigation for unmanned helicopters

Abstract

The development of a vision system to aid the autonomous navigation of an unmanned helicopter, primarily based on inertial sensors and GPS data, is presented. An unmanned helicopter has been equipped with appropriate sensors and a vision system fitted on a custom pan-tilt mechanism. Robust software, based on the Open Computer Vision Library (OpenCV), has been developed for handling images and video from a camera. Our implementation involves real time object recognition, histogram matching for real time video streaming, pattern matching and object tracking. Software implemented in C++ interacts with Matlab in order to aid the autonomous navigation of the helicopter.