A METHOD FOR DETERMINING THE COORDINATES OF OBSERVED OBJECTS BASED ON THEIR SEMANTIC OR ASSOCIATIVE FEATURES BY THE TELEVISION SYSTEM OF AN UNMANNED AERIAL VEHICLE

Abstract

A method for determining the coordinates of observed objects based on their semantic or associative characteristics by a television system of an unmanned aerial vehicle, which differs from known methods in the step-by-step implementation of the following procedures: The first stage is the selection of the data processing level in the infological representation based on the cognitively formed request, flight task, mission, operator of the unmanned aerial system; at the semantic level, data processing is carried out using one of several recording devices; at the associative-ontological level, data processing from all recording devices is possible; The second stage is, at the semantic level, detection, recognition, and identification of observed objects using pre-trained convolutional neural network algorithms, and at the associative-ontological level, intra-ontological or inter-ontological recognition and identification of objects using pre-trained multilayer perceptrons; The third stage is the selection and calculation of rectangular windows of identified objects based on their semantic or associative features; The fourth stage is a comparison of the sizes of rectangular windows of the identified objects with a pre-formed empirical database of linear dimensions; The fifth stage is the calculation of the range to ground-based objects of observation by a television system relative to a small-sized, low-altitude, low-mobility aircraft; The sixth stage is the calculation of the coordinates of objects from the formed set based on range, angles of the center of the windows of the detected objects relative to the orientation of the UAV in space, and global coordinates of the aircraft.