Wheat yield estimation using UAV-imagery in the arid climate of Transbaikalia

The improvement of the scientific bases of unmanned monitoring of agro-systems in different natural and climatic conditions represents a crucial task for the advancement of this field of study. Wheat is a primary crop in the structure of crop production. This study aims to estimate wheat yield in arid conditions of Transbaikalia with the use of vegetation indices from UAVs and modeling by neural networks. The findings indicate that yield prediction is feasible up to the milk ripeness phase. Spectral vegetation indices demonstrate a correlation with yield that is slightly higher than that of colored vegetation indices. Neural network modeling of wheat yields generally exhibits a lower performance than linear regression.