Multi-layer neural network for solving problems of recognition of drained crop areas

Abstract

The article deals with the design and software implementation of neural network modules for solving problems of image recognition. In particular, it describes the development of a module for network training and recognition of input pulses, which made it possible to recognize, process and analyse aerial photographs of agricultural crops as objects of identification based on the use of a multi-layer deep learning neural network. The practical use of the software tool is possible for the study and study of the peculiarities of cultivation in conditions of irrigation and differentiated placement.