Agriculture industry in Russia is an important part of the economy. Therefore, this industry must be highly efficient and profitable. The usage of remote sensing data for field monitoring makes it possible to simplify the process of collecting information on sown areas and increase the efficiency of this economic sector. Agricultural crops show up well on satellite images and are relatively easy to decipher both in terms of texture and spectral characteristics. However, even for deciphering well-distinguishable objects, large labor costs and high qualification of the performer are required. Recently, active work has been carried out to automate monitoring performed on satellite images, however, these methods are at the initial stage of development and are mostly created for narrowly focused individual areas of activity. The aim of the study is the development of a methodology for automated processing of spatial data of agricultural designation. The novelty of the research lies in the absence of methods for automated decoding in this area. The article describes a study of the possibilities of automated image processing, the analysis of the necessary spectral characteristics and vegetation indices, and the application of the machine learning algorithm used in the methodology. The development of software and hardware implementation of the methodology in Python programming language for geographic information system QGIS is presented. The software and hardware implementation of this methodology allowed to perform automated processing of images of agricultural designation and to identify the objects displayed on them.