Visual Positioning Method for Unmanned Aerial Vehicle Charging Platform Using Cooperative Target

This paper proposes a precise landing visual positioning scheme for an unmanned aerial vehicle (UAV) charging platform. By using a two-level cooperative target and feature segmentation method, the perimeter and area of the connected domain are calculated for preliminary fast rough segmentation, and then precise segmentation is performed based on the geometric features of the connected domains, and the goal of quickly filtering a large number of connected domains is achieved. Through real-time testing with an airborne pixel camera, the average processing speed reaches 35.3 ms, the problems of high-precision positioning and real-time performance in the process of UAV automatic landing have been solved. Experiments have shown that proposed visual positioning scheme can improve the efficiency of UAV automatic landing, and provide the good practical value for the application and promotion of fully automatic UAV inspection.