Methodology for assessing the control system of aircraft weapons in the process of aiming controlled aircraft weapons equipped with opto-electronic homing heads

This article discusses the increase in the effectiveness of the combat use of an air-to-surface guided missile based on the received target designations from an unmanned aerial vehicle, which makes it possible to automatically solve the tasks of searching, detecting and recognizing targets in real time conditions at large distances in a complex background-target situation in the combat zone actions, thereby reducing the risk of the carrier aircraft falling into the enemy air defense coverage area. An optimized architecture of a convolutional neural network has been developed for image segmentation and ground target recognition in the optoelectronic system of an unmanned aerial vehicle, as well as an algorithm for automatic recognition of a ground target by an artificial neural network in a television homing head of controlled air-to-surface weapons. An analytical comparative study was carried out on the probability of hitting a ground target such as a tank between the developed algorithms for automatic recognition of a ground target and the use of visual (optical) detection and recognition of a ground target by a pilot (navigator) at different values of the average intensity of the flow of fire from enemy air defense missiles. The software implementation of algorithms for automatic recognition of a ground target and training of an optimized neural network using the object-oriented programming language Matlab has been implemented.