Dynamic monitoring of land use/cover changes based on CBERS imagery

The work present in this paper focuses on using CBERS imagery data to acquire some information such as land use/cover change (LUCC) by means of remote sensing and GIS and to find out an appropriate classification method to CBERS imagery by applying and comparing some classification methods. CBERS CCD imageries of Xuzhou city in 2001, 2005 and 2008 were classified by using decision tree classifier, support vector machine, maximum likelihood classifier and neural network classifier. After comparing the applicability and accuracy of four classification methods, the data classified by decision tree classifier was chosen to extract the land use information. In order to monitor the land-use, land-use dynamic degree and land-use preference index were introduced. The dynamic changes of land use and land cover from 2001 to 2008 were analyzed by utilizing GIS. Recent experiments in Xuzhou city suggested that the area was expanded rapidly and its construction area was being extended continuously. This research we have done suggests that the CBERS data has good application prospective in land use monitoring.