Processing of the information by the complex of neural networks in the distributed fiber-optical measuring systems

The paper discusses tomography reconstruction of distributed physical fields by means of distributed fiber optical measuring systems (FOMN) for incomplete parallel schemes of measuring lines (ML) stacking. The approach is presented, consists in optimization of geometry of a measuring network for the purpose of the further application neural network methods of restoration of a full-image of investigated functions. Possibility of a choice and use of a suitable neural network from set of the several, in advance trained, neural networks of RBF-type is investigated.