Edge-based target detection for unmanned aerial vehicles using competitive Bird Swarm Algorithm

Target detection for unmanned aerial vehicles is an important issue in autonomous formation flight. In this paper, a novel target detection approach for unmanned aerial vehicle formation is proposed based on edge matching. The windowed edge potential function is utilized to describe the attraction field for similar edges. Afterwards, the edge-based target detection problem can be formulated as an optimization problem. An improved version of the bird swarm algorithm, which is called competitive bird swarm algorithm, is proposed to find the location, rotation angle and scale of a given template on a specific image. A strategy named “disturbing the local optimum” is designed to help the original bird swarm algorithm converge to the global optimal solution faster and more stably. Formation flight platforms, which consists of unmanned aerial vehicles moving in leader-follower pattern, are used in our experiments. Images obtained by vision sensors embedded in the leaders are used to verify the effectiveness of the proposed method. The proposed algorithm is tested on both indoor and outdoor images to demonstrate the robustness. Comparative experiments with other state-of-the-art algorithms, including genetic algorithm, particle swarm optimization, artificial bee colony algorithm, pigeon-inspired optimization, and the basic bird swarm algorithm, are also conducted. The results prove the superiority and robustness of the proposed target detection algorithm.