Abstract
The article provides a general overview of tethered multicopters, describes their main functional capabilities, and formulates the task of developing an autonomous navigation system for use in one of the most challenging operational processes of tethered unmanned aerial vehicles (UAVs) - their landing on the ground. The use of keypoint technology is proposed for the development of the autonomous navigation system. A brief review of keypoint detectors and descriptors is presented. It is shown that the FAST detector can be used as a keypoint detector. However, for descriptor construction, a mathematical model based on the Ψ-transformation method used in nonlinear function optimization problems was employed, forming the basis for the development of a proprietary keypoint descriptor - the Ψ-descriptor. The results of the practical application of the developed Ψ-descriptor for solving UAV autonomous navigation tasks are presented. The proposed new method for keypoint determination can be used in the computer vision systems of tethered unmanned aerial vehicles to solve landing tasks as an alternative to conventional GPS-based systems.