ESTIMATION OF PLANTS HEALTH USING CONVOLUTIONAL NEURAL NETWORKS

Abstract

The article continues the authors’ research in the field of building systems for automatic diagnostics conditions and dynamics of plant growth. New concepts of intelligent farming, where field conditions are controlled by autonomous systems, have now become widely used. Building such systems requires solving several problems related to digital image processing, such as detecting plants and identifying their condition. Therefore, the article studies the actual problem of constructing an automatic system for assessing the state and dynamics of plant growth. As the central component of the system under consideration, the authors propose to use the mathematical apparatus of convolutional neural networks. The article describes the architecture, training procedure, and testing of the neural network. The results of computer experiments are presented. The analysis of the work of convolutional neural networks shows the high efficiency of the proposed solution to the problem of assessing the state and growth dynamics of biological crops.