Algorithm of random distribution of MSTAR images on a radar background in order to increase the efficiency of information processing

An innovative algorithm is proposed for generating realistic radar images by randomly distributing objects from the MSTAR library onto synthesized backgrounds, which solves the problem of the lack of open source data for military purposes and their high cost of obtaining. The algorithm integrates statistical speckle noise matching to harmonize the characteristics of the target and background, unsubdiscretized contour transformation (NSCT) to preserve structural details during fusion, as well as a Lie filter to reduce noise during the post-processing stage. The method allows you to generate thousands of realistic images simulating radar shooting conditions (angles of incidence, polarization), which improves the training of AI models. The results demonstrate increased classification accuracy compared to traditional methods in compliance with SAR standards. The research contributes to the development of computer vision for security tasks by offering a solution in conditions of limited data.