Construction perspectives of the remote sensing data high-performance processing system

The paper presents the intermediate results of studies on the challenges of full color and multispectral remote sensing images processing. Presented the experiments results of target regions detection with spectrographic fill and neural network clustering; solved the problem of complex rigid objects detection. Proposed implementation of the multispectral images to grayscale and color algorithm conversion. All programs are designed for operation on a heterogeneous computing machine. (In Russian).