Parallel Classification of Large Aerospace Images by the Multi-alternative Discrete Accumulation Method

Abstract

Abstract. The paper deals with parallel large aerospace images processing. We considered a simple multi-alternative discrete accumulation method for reliable distinction of satellite imagery and implemented a parallel classification system to increase the algorithm efficiency. The process of development of the distinction algorithm and system architecture was described. The system prototype was successfully tested. The experiments allowed to draw conclusion about the system performance and to estimate the effect of using the parallel architecture. The considered approach could be used in complex neural networks processing.