The comparison of change detection methods in remote sensing

Abstract

Recently, It is issues the master plan about land-monitoring system construction. The land monitoring support to making decision efficiently, and to management resources. The change detection technology is good to detection changed area using satellite images for land monitoring. There are some summary researches of various change detection methods about strength and weakness already. But they didn't worked verification process by test area. In this study, to verify what is the best of change detection methods in remote sensing, Through to compare the change detection results using Landsat 7 ETM+ images. In accuracy assessment process, the kappa coefficient is estimated by comparison relative with each change detection methods. It is different origin computing method of kappa coefficient. The results shows the image ratioing by NDVI method is the best to detect changed area. And the image differencing by PCA method is good to detect changed area. Additional the image differencing by tasseled cap is very similar to the image differencing by principal component analysis method.