The paper introduces how multi-class and single-class problems of searching and classifying target objects in remote sensing images of the Earth are solved. To improve the recognition efficiency, the preparation tools for training samples, optimal configuration and use of deep learning neural networks using high-performance computing technologies have been developed. Two types of CNN were used to process ERS images: a convolutional neural network from the nnForge library and a network of the Darknet type. A comparative analysis of the results is obtained. The research showed that the capabilities of convolutional neural networks allow solving simultaneously the problems of searching (localizing) and recognizing objects in ERS images with high accuracy and completeness.

меТоД СИНТеза улуЧшеННого ИзоБРаЖеНИя По СеРИИ СНИмкоВ ДИСТаНЦИоННого зоНДИРоВаНИя землИ 1 куракин а.В., 2 абрамов Н.С., 2 карпеш С.В., 2,3 хачумов В.м.

No abstract