Robust marker-tracking system for vision-based autonomous landing of VTOL UAVs

摘要

This article presents the development of a visionbased VTOL UAV docking system for outdoor applications. A new marker type is constructed, along with a robust detection algorithm based on the Hough transform, which forms the basis of the positioning system. For verification purposes, a binary classifier based on support vector machines is trained. Test data show great robustness against varying lighting, partial occlusion and unsharpness. The practical usefulness is demonstrated by creating a single tracking module, based upon the PX4FLOW camera. Equipping a quadrotor with the new positioning sensor allows it to perform a precision landing.