Development of a method for determining the parameters of the segmentation process and ground target pattern recognitionin opto-electronic homing heads

The process of image segmentation and recognition of ground targets in a opto-electronic homing head of guided aircraft missiles is considered. A technique has been developed for determining the parameters of the process of segmentation and pattern recognition of ground targets and a block diagram of an artificial neural network, optimized by the number of neurons in the hidden layer, for recognizing typical ground targets. Studies of segmentation of frames (images) from a video sequence were carried out using the developed algorithm for segmenting images with a complex background, followed by the implementation of the process of pattern recognition of typical ground targets. A software implementation of the algorithm for segmentation and training of a neural network has been carried out using the object-oriented programming language MatLab & Simulink.