Results of using geoinformation and statistical analysis methods to study spectral reflectance characteristics of agricultural crops of Belarus

Abstract

The results of using geoinformation and statistical analysis methods to study spectral reflectance characteristics of the nine most typical agricultural crops of Belarus are presented. Spectral brightness coefficients and normalised difference vegetation index (NDVI) values were extracted from Landsat-8 multispectral satellite images in the software package ENVI (version 5.2) and analysed based on the methods of zonal statistics in the software complex ArcGIS (version 10.2) and mathematical and statistical analysis in the program Statistica (version 10). The verification of satellite data with the corresponding field measurements was carried out on the basis of correlation analysis, namely, a reliable strong positivelinear relationship between the measured in the field by a specialised GreenSeeker instrument NDVI values and the calculated by Landsat-8 satellite data NDVI values was established. The character of the distribution of spectral brightness coefficients and average NDVI values depending on the type of agricultural crop was assessed using a dispersion analysis, which allowed revealing patterns hidden in the spectral data. In particular, after applying the procedure of multiple comparisons using post hoc tests, it was established which types of crops significantly differ from each other and for which dates these differences were observed. The obtained scientific results were systematised and presented in the form of correspondingtables. The data contained in the tables made it possible to improve the methodology of automated recognition of the crops considered in the study.