Selecting the Visual Landing Markers Located on a Ship Deck for UAV Automatic Landing

摘要

Of all the flight modes of unmanned aerial vehicles (UAVs), the landing approach and the landing itself are the most difficult and dangerous ones. To perform an automatic landing on a ship deck using a computer vision system (CVS), a UAV must analyze the visual landing markers located on the deck. When the UAV is about to start landing, these markers must be clearly visible and discernible. The aim of this study is to detect markers in order to build a trajectory taking the UAV to an imaginary boundary of a heterogeneous group of ships that includes the target ship. The cutting-edge method of solving this problem consists in running tests identifying the optimal conditions for a successful search for and recognition of landing markers installed on a ship deck. The experiment covers eight types of markers. The research findings back the efficiency of the developed technology for the computer vision-aided detection by UAVs of visual markers located on a ship deck. Of the eight possible types of landing sites, the one was chosen that demonstrated the best detection and recognition capabilities. Research has proved the efficiency of the proposed algorithms for solving this problem. Unmanned aerial vehicles were used as a tool for designing and building a system for the automatic detection of artificially embedded markers on images supplied by the UAV on-board electro-optical system.