With the rapid progress in the development of information technologies and methods of remote sensing of the Earth the computational capabilities and the volume of initial information are significantly
expanding. As a result, the problem of processing high-resolution aerospace images arises. This problem is associated with data redundancy, when the plot 1 ha corresponds to 4 million pixels. In this
regard, it has been proposed to initially reduce and determine the optimal amount of high-resolution
image data required to solve the issues of precision agriculture, in order to avoid time-consuming computations and increase the calculation efficiency. The paper presents an approach to assessing the feasibility of the transition to variable-rate agrochemical application technologies. The proposed approach
is based on a variogram analysis of the within-field variability of the optical characteristics of plants.
The results show, that for the imagery with a resolution of 10 cm per pixel is the most appropriate to
take into account only 0.5–1 % of the total number of pixels (with a uniform distribution of points in
the imagery). The presented approach can be used in other directions related to the geostatistical analysis of optical indicators calculated from a particular set of pixels depending on the spatial resolution of
aerospace images.