Backscattering coefficient measurement and land use land cover classification using ENVI SAT ASAR data

Abstract

The polarimetric SAR data of the space borne sensor, ENVISAT-ASAR (Environmental Satellite - Academic & Science Astronomy & Space Science) has been used for the land use land cover classification of the study area. It was an earth observing satellite operated by the European Space Agency (ESA). Its mission was to observe the earth and monitor critical aspects of the environment such as climatic changes on the earth at the local, regional and global levels. The data set of this sensor is a dual co-polarization amplitude data consisting of HH and VV channels. Initially various incidence angle images such as sigma naught, beta naught and gamma naught have been generated for both HH and VV polarizations. Then the backscattering coefficients of different features such as water, bare soil, vegetation and urban have been calculated. The backscattering coefficient values of the HH polarization are high compared to the values that are obtained with VV polarization. Then the land use land cover classification has been done by implementing different supervised classification algorithms. These classification methods are Parallelepiped, Minimum Distance, Mahalanobis, Maximum Likelihood, Binary Coding and Support Vector Machine. Then the accuracy measurements have been done for all these classification methods. In the present study the accuracy results obtained with the supervised Support Vector Machine classification algorithm are more compared to the accuracy results obtained with the other supervised classification methods.