Technology development for detecting inhomogeneities in agricultural fields

Abstract

The paper addresses the relevant problem of identifying inhomogeneities in crop development based on analysis of multispectral images obtained from Earth remote sensing (ERS) satellites and unmanned aerial vehicles. Various techniques for detection of inhomogeneities are considered, which are based on computation of normalized difference vegetation index and its analysis, but do not take into account stages of crop production. The authors propose an original method and new algorithms for detecting inhomogeneities that take into account field zoning at various stages of the plant development cycle: preparing the field for sowing, emergence, and development of seedlings, heading. The paper also describes comparative analysis of results of the algorithms described in this paper and those of system-analogs. This analysis confirms effectiveness of the proposed algorithms. The method and algorithms for detection of inhomogeneities, developed by the authors, are used in the software module of image processing and presentation of ERS results for solving problems of precision farming, providing prompt, flexible, and efficient results for consumers.