A method for unmanned rotorcraft to detect circular markers

Detection of circular markers in a complex environment is an important task for the unmanned rotorcraft. In this paper, according to the analysis and comparison of computational complexity, noise sensitivity and detection accuracy between the standard Hough transform (SHT) and the gradient Hough transform (GHT) algorithm, it is proposed a hybrid algorithm for circular markers detection (HA-CMD) which balances the constraint of computational capability of the embedded system and the requirement of detection accuracy. Corresponding to the altitude, there are two strategies utilized, that, the adaptive radius algorithm combined with SHT and cluster analysis is applied in the high-altitude recognition area, moreover, the former circular tracking technique combined with GHT is introduced to reduce misjudgment in the low-altitude area. The proposed algorithm is tested on an unmanned rotorcraft. The experiment results demonstrate the effectiveness and show an effectively balance.