Long-term tracking method on ground moving target of UAV

Abstract

The 7th IARC (international aerial robot competition) focus on a long-term and real-time tracking of ground moving target for UAV (unmanned aerial vehicle). This paper proposes a tracking method using Camshift algorithm with color and edge features fusion, which will apply a Phash target detection algorithm basing on perceptual hash algorithm. Supposing targets are severely obscured or disappear from view, which will lead tracking failure, whereas targets can be detected and obtained again using the method. Experimental results demonstrate that this method can provide a long-term and real-time tracking of ground moving target, which processes both robustness and computational efficiency.