THE USE OF ARTIFICIAL NEURAL NETWORKS TO IMPROVE THE RESOLUTION ACCURACY OF REMOTE SENSING INFORMATION ON THE EXAMPLE OF LEGUME CROP CULTIVATION IN EASTERN KAZAKHSTAN

Abstract

Innovative development of agro-industrial complex today is associated with the implementation of Earth remote sensing (ERS) data. This study presents the development of a mathematical model using artificial neural networks to improve the accuracy of remote sensing information resolution on the example of growing grain legume crop peas in the soil and climatic zone of eastern Kazakhstan. The model is based on artificial neural networks using specialized libraries TensorFlow, Keras based on calculations of correlation coefficients between the values of spectral indices obtained from satellite imagery processing and unmanned aerial vehicle (UAV) data. The results demonstrate the potential of using ERS data and artificial neural networks to improve the resolution of space imagery information in crop cultivation, providing more accurate and timely information for decision-making in smart agriculture.