Extracting Land use Information from Beijing-1 Micro-Satellite Images

Extracting landuse information from Beijing-1 micro satellite images in Longqiao town, Chengdu city, China is discussed here. Firstly, the images are geometrically corrected and registered by several ground control points. Secondly, the images such as panchromatic band with the resolution of 4 meters and multi-spectral band with the resolution of 32 meters are enhanced and fused. Thirdly, the Maximum Likelihood Classifier (MLC) is used to extract landuse information from the fused images. Finally, the Method based on Knowledge Discovering (MKD) is used to extracting landuse information from the Beijing-1 Micro-satellite images. The accuracies of the landuse information extracted by the two methods are evaluated by sampling and visual interpretation. The average accuracies of MLC and MKD are 65.7% and 89.3% respectively. It is shown that the landuse information with higher accuracy can be obtained from the Beijing-1 micro-satellite images by using MKD.