Application Research of Support Vector Machine in Multi-Spectra Remote Sensing Image Classification

Abstract

In order to improve the accuracy of multi-spectra remote sensing image classification, a terrain classification method based on support vector machine is proposed. A remote sensing image classification method based on SVM algorithm of C-SVC type is introduced and emphasis is put on the study of the improved SMO algorithm. In order to improve efficiency of classification, multiple-spectra remote sensing image classification of terrain is classified using fuzzy clustering based on fuzzy c-means algorithm which adopt semi-supervised improved algorithm. The experimental results show that the approach has an advantage over traditional classification methods.