Deriving Urbanization Pattern of Guangzhou City with Remote Sensing Imagery with Decision Tree

Abstract

In this paper, urbanization pattern of Guangzhou city is studied with land use/cover maps estimated from 11 Landsat TM imageries with an interval of 2 to 4 years by using C4.5 decision tree algorithm. It has been found that since 1987, the urban area has expanded twice which are mainly converted from cultivated land and the expanding is mostly around the urbanized area. Based on the relationship among social economy, population and land use change, the driving forces for urbanization are different in different periods, including national policy, the aggregation and diffusion of population and the economic development.