Neural network system for recognition and visualization of problem areas of crops in precision farming: justification of application and aspects of implementation

Abstract

Abstract The problem of recognizing the state and visualization of problem areas of crops in precision farming includes agrobiological, optical-spectral, mathematical-algorithmic and other aspects. The difficulties arising in the process of analyzing and classifying images of remote monitoring of crops necessitate the development and adaptation of methods of their computer analysis using neural network algorithms. The results of the study showed that the most effective approach to image classification is a combination of preprocessing methods followed by recognition by neural network algorithms for the rapid identification of defective areas of various nature.