Classification of High Spatial Resolution Remote Sensing Image Using SVM and Local Spatial Statistics Getis-Ord Gi

摘要

In this paper, the support vector machine (SVM) algorithm was utilized to tackle the classification of high resolution images from airborne digital sensor systems. Firstly, the original image was classified using SVM of four common types of kernel functions, namely linear, polynomial, RBF and sigmoid function, and the SVM with RBF kernel function can achieve the most satisfactory result. On the other hand, Getis-Ord Gi, one type of local spatial statistics, had been calculated with varying lags from 1 to 10. When classifying Gi image with lag of 3 using SVM of the RBF kernel function, an overall accuracy of 95.66% was achieved, which is more satisfactory than the result from the original image. The result shows that Gi images with lags less than the variogram range can be used instead of the original multi-spectral image to improve classification accuracy between features with similar spectral characteristics like trees and lawns, as a result, to increase the overall classification accuracy.