ALGORİTHMS FOR FİNDİNG NON-İNTERSECTİNG ROADS ON İMAGES

Abstract

Research relevance The use of GPS for UAVs to determine their current coordinates (navigation) leads to a number of problems. The development and improvement of the autonomous navigation method without the use of GPS is an actual issue, and this issue is considered in the article. The subject of study in the article is the issue of UAVs navigation based on video images. The purpose of work to develop an algorithm for autonomous navigation of UAVs that requires less resources. The following tasks are solved in the article: software modules are developed based on image processing algorithms; suitable model for invariant identification of curves was studied and applied. The following results were obtained: an algorithm was developed for the invariant identification of winding roads (curves) without intersections from different angles; developed software modules have been tested in the Republic of Azerbaijan on the basis of a map of the region called "Agsu Pass". Conclusions: the application of the method proposed in the article will enable rotation-invariant autonomous navigation of UAVs using less resources.