Development of An Algorithm for Data Augmentation in The Problem of Object Classification on SAR Images using Convolutional Neural Networks

Convolutional Neural Networks have achieved great success in optical image processing tasks. There are many open data sets that can be used when creating your own model to solve any problems. However, if the problem is related to images obtained using a Synthetic Aperture Radar, then the number of open data sets is very limited. This article explores the problem of using Convolutional Neural Networks to classify objects in SAR images using a limited dataset. An algorithm for augmentation of radar images is presented. The possibility of a significant increase in the accuracy of object classification is shown, due to the multiple increase in the data set by unique images of the studied objects.