A COMPARATIVE ANALYSIS OF UNIMODAL AND MULTIMODAL DEEP LEARNING METHODS FOR THE EARLY DETECTION OF CROP DISEASES

The thesis presents a comparative analysis of unimodal (image-only) and multimodal (image, spectral, sensor, and meteorological data based) deep learning approaches for the early detection of crop diseases. The advantages and limitations of both directions are systematized by criteria, and the superiority of the multimodal approach under field conditions is substantiated.