Remote Sensing Bid-Data Classification with Support Vector Machine

Abstract

Remote sensing is the process of getting knowledge regarding some article or observable fact not includingcreationmental contact with the object. The data collected by deploying this method is termed as the remote sensing data. Data collected by this method may be either linear or non-liner in nature. For classification of linear statistics, we haveused linear Support Vector Machine (LSVM) and for non-linear Support Vector Machine (NSVM) using different types of kernels.

Use of LSVM offers higher accuracy as compared with NSVM. In this paper, we have implemented concept of SVSA (Support Vector Selection and Adaption) for non-linear data with implementation, we have observed that this method offers higher accuracy as compared to selecting different kernel functions.We will use RACE data for training purpose, which will extent that the result of classification using this method which by passes the result of LSVM.