Vision-Based Object Tracking Algorithm With AR. Drone

摘要

This paper presents a simple and effective vision-based algorithm for autonomous object tracking of a low-cost AR.Drone quadrotor for moving ground and flying targets. The Open-CV is used for computer vision to estimate the position of the object considering the environmental lighting effect. This is also an off-board control as the visual tracking and control process are performed in the laptop with the help of Wi-Fi link. The information obtained from vision algorithm is used to control roll angle and pitch angle of the drone in the case using bottom camera, and to control yaw angle and altitude of the drone when the front camera is used as vision sensor. The experimental results from real tests are presented.