Identification of deviations of unmanned aerial vehicles by parallax images

Abstract

The article is devoted to the study of the identification of deviations of unmanned aerial vehicles from overlapping images of the underlying surface without determining the parameters of their mutual orientation. For this purpose, a mathematical model has been developed that establishes a functional relationship between the parallax differences of overlapping images and the deviation parameters of unmanned aerial vehicles. In this case, all possible types of deviations and options for their combinations are taken into account. Studies have been carried out and estimates of the influence of deviations on the magnitude of the change in the parallaxes of overlapping images have been obtained. The results obtained allowed us to conclude that during the flight of unmanned aerial vehicles it is possible to automatically identify and evaluate their allowable deviations from a given trajectory using the parallaxes of overlapping images without determining the parameters of their mutual orientation.