Vegetation change detection based on IRS and Landsat Satellites data

摘要

The study aims to evaluate the IRS-P6 LISS III and Landsat ETM+ efficiency in plant group identification. In order to achieve this purpose, 143 training samples were collected from a homogenous plant species composition with an area of 3600 m~2 (60 x60 m). Coordinates of these training samples recorded using GPS and transferred to a GIS environment. For satellite data ENVI 4.2 software has used to process and analyses them. Several methods of processing such as; spectral separability, supervised classification and classification accuracy assessment have used in order to gain a satisfy evaluation accuracy. The results of this process indicated that the best separability is related to net farming of Me.sativa and Ju.poly carpus -Ar.kopetdaghensis community (1.99 for Landsat data and 2 for IRS). In contrast, the worst results were related to Ju.-polycanpus-On.comuta and Ju.polycarpus-Ar.kopetdaghensis communities (1.57) for Landsat and Ju.polycarpus-Ar.kopetdaghensis and Ju.poly carpus-Ag. intermedium communities (1.53) for IRS data. It can be concluded that the satellite data are roughly able to identify plant groups when vegetation communities have a sufficient homogenous, wealthy and ecologically separable zones.