Application of the Method of Statistical Differentiation for the Gridding of Aerophoto Image to a Satellite Image

Abstract

The work deals with the creation of an autonomous navigation system for unmanned aerial vehicles by way of gridding airborne images to a satellite image of the terrain. This system can be used for navigation in conditions of global positioning signal (GPS, GLOBASS) interference while maintaining the radio silence mode. Special points combined into triangles are calculated according to the images of the airborne image of the aircraft and the corresponding part of the satellite image. Each triangle is described by a descriptor, which includes information about the types of special points forming this triangle, the order of their visiting, and the position of the vertex of the triangle relative to the base. Using the method of statistical differentiation can significantly reduce the amount of enumeration by comparing only equivalent triangles with the same descriptors. The result of comparing the triangles of the two images is the set of pairs of corresponding segments, among which the pair with the best result of correlation criteria is selected. The chosen pair of segments gives the parameters of the shift, rotation and the scaling factor, which allow you to tie the coordinates of the picture of the onboard camera of the aircraft to the space image of the terrain. In the conclusion of the paper, an example is given of the constructed trajectory of an unmanned aerial vehicle with reference to a satellite image.