论文部分内容阅读
Impervious surface is one of the important parameters of valley water circular simulation, scientific estimation for which has significant and practical value for the urban water quantity and process simulation, diffuse pollution estimating and the forecast of climate changes. The objective of this research is to get the information of impervious surface and its dynamic change. Through the computer-assisted field method, the technologies of decision tree and data mining were applied to withdraw the impervious surface information in research region by the Landsat TM data in 1988, 1994 and 2002. The results suggested that the accuracy of impervious surface information extraction in the study area arrived above 94.4% in 2002 image. On this basis, the mixed method was used to extract the location and the types of the impervious surface change. The overall accuracy of monitoring reached 89%, which meets the demand of the hydrological models.