Abstract:
This article is devoted to a set of important areas of research: the analysis of formal
representations and verification of pests and pathogens affecting crops using spectral brightness coefficients
(SBR) for the period from 2021 to 2023. The database contains about 10,000 records covering the growing
season, types of diseases and pests, as well as their growth phases in a real coordinate system. The work
uses machine learning techniques including logistic regression, extreme gradient boosting (XGBoost), and
Vanilla convolutional neural network (CNN) to analyze spectral data and classify the presence of pests and
diseases in satellite images. The main goal of the work is to optimize and improve the quality of agricultural
productivity through early detection and accurate classification of pests and diseases in the agricultural
sector. The results of the study can be applied in the development of innovative agricultural systems that
will increase yields, reduce the cost of pest and disease control, and optimize production processes. The
conclusions of this work can be used both as scientific and practical recommendations for agricultural
enterprises and organizations and for the development of new technologies and programs for automating
agricultural processes. The use of machine learning techniques and spectral data analysis promises significant
breakthroughs in the agricultural sector, helping to improve the efficiency, sustainability, and quality of crop
production.
