The decomposition of the deep learning machine for specialized datasets for time minimizing of spatial information processing

For effective military management, the creation of complexes automation tools, spatial information processing systems, primarily consolidated, is a priority task in conditions constant growth data and requirements for their collection, transmission, storage, use.
 The problem consolidated related to diversity sources, formats, acquisition use, etc. All that imply an extremely complex organizational technical structure, kind ‘system systems’.
 Deep Learning Machine (DLM) ensure high accuracy prediction. But such DLM should be matched usage where time restrictions space lack are present.
 So, effort create dataset maximum rests on computing resources (in common research it possible overcome but sphere isn’t). In applied tasks, criterion overcoming uncertainties due confrontation critical. This allows us put forward hypothesis impossible achieve absolute deep learning machine. Therefore, variable tactical situations, advisable specialized datasets efficiency from each iteration step (or combination) using decomposition method system. It analysed methods scientific solutions machine systematizing types existing situations. end detection recognition system with set proposed paper.
 volume at level 103 enables ultra-high speed processes person up without excessive demands. itself revealed.
 principle forming dataset, or sets, obtaining high-accurate -fast systems.