Parallel program system for HIL experiments on visual navigation of unmanned aerial vehicles

The article is devoted to the development of a program system designed to solve the navigation problem for unmanned aerial vehicles using the methods and algorithms of computer vision, image processing and analysis. The system operates on a cluster computer. The sources of data for solving the problem are the visual navigation are the HIL data — the results of generation of video sequences from virtual UAV. A subsystem of flight simulation and video generation is developed. Also the algorithms for solving the problem of visual navigation of UAV in flight over flat terrain and terrain with relief are developed. The results of experiments demonstrating the effectiveness of cluster computer in the problem of preliminary processing of reference images of terrain for subsequent visual navigation solutions, as well as in the problem of parallel processing of multiple independent video sequences coming from different UAVs are presented. (In Russian)