Tích hợp ảnh radar và ảnh quang học xây dựng bản đồ hiện trạng sử dụng đất thành phố Cần Thơ

Abstract

Land use maps play an important role in statistics, inventory and orientation of land use planning. In recent years, the application of remote sensing data has been contributing to the generation of land use and land cover maps effectively. However, optical image is often covered by clouds, especially in the rainy season, which affects the accuracy of classification results. With the characteristics of radar images that can penetrate to clouds and weather conditions, the study used fusion method of two types of optical (Landsat-8) and radar (Sentinel-1) images in generating land use map of Can Tho city in 2019. Using Gram-Schmidt method to integrate two images and applied object-based classification approach to generate the land use map of Can Tho in 2019. The classification result showed that there were 5 main land uses in Can Tho city: perennial crop, paddy rice, canals/rivers, settlement and aquaculture land. The results of accuracy assessment showed that the total accuracy was 83.8% and the Kappa coefficient was 0.68. The results of the study showed that there was great potential of integrating radar and optical imagery in mapping land use map Can Tho city.