Integration and Implementation of a Low-cost and Vision-based UAV Tracking System

We present in this paper some preliminary research work carried out for SheLion, a vision-based unmanned aerial vehicle (UAV) designed and built by the UAV research team at the National University of Singapore. Unlike HeLion, a twin brother of SheLion, the latter is equipped with an onboard camera and image processing system, which is capable of detecting and tracking ground targets. We first give in this work a brief introduction to the UAV system, which consists of a bare helicopter, an onboard flight system and a ground supporting unit. We then focus on the development of a real-time vision algorithm, which is able to efficiently detect and identify ground targets. Experimental result based on images captured in actual flight tests is also presented.