Comparison Analysis Of Land Use Classification Between ASTER And ETM+ Remotely Sensed Image

With the development of remote sensing technique, multiplatform multitemporal, multi-scale and higher spectral revolution remotely sensed data used more and more domain on the application in geosciences. Classification of land use is the base study on the land resources, land evolution and land use/cover change. Landsat-7 Enhanced Thematic Mapper data is the most used remotely sensed image. Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) on TERRA have more spectral and spatial resolution in VNIR.This research comparison the ASTER and ETM+ data used in the classification of the land use of Guangzhou Huadou. According to the national land use classification criteria and the characteristic of study area, classify the land use for seven types as urban, forest, water, pond, garden, agriculture and bare land. Maximum likelihood classifier (MLC) is the most used and effective classification method. First, classify the land use of Huadou city by MLC using ASTER and ETM+ image data supervised classification. Their precision is 0.91 in ASTER and 0.89 in ETM+. Then analyze the effect of the classification in the eight towns in Huadou and the seven land use classes. Main Conclusion include the ASTER has the higher accuracy in most towns and types. But ETM+ plays much better in forest class and bare land class. The classification that the confusion that the effectiveness that has much better conclusion.At last, this research also analyze the results overall precision and Kappa index by different classify method and the land use/cover change from 1988-2003.