Abstract

Land cover change detection is the key point in global change researching and remote sensing analysis is an important technological method for change detection. In this paper, a land cover change detection method based on texture features is proposed by using old digital land cover vector map, old remote sensing image and new remote sensing image of areas studied. Firstly, the two remote sensing images are segmented by old land cover vectors, and texture features are acquired through co-occurrence matrix. Secondly, Euclidean distance is calculated to judge whether land covers change or not. Finally, Gabor filters is adopted to calculate parameters of each pixel in those areas where changes take place. Thus we can confirm which kind of land cover type changes and locate where changes take place. Through accuracy evaluation, the method proposed performs well for the instance discussed.

Paper Details