关键词: 遥感, 城市扩张, 决策树方法, 形状指数, 环杭州湾地区

Abstract: Taking the region around Hangzhou Bay as a study area, the urban land information was extracted from TM images of 2003 and 2006 through the decision tree method together with the shape index. And the characteristics of urban expansion in cities of Hangzhou, Shaoxing and Ningbo were further discussed. The overall Kappa coefficients of urban land were all above 0.87 through precision test and proved that: it was feasible to extract the information of the urban land. The area of urban land in 2003 and 2006 was 914.04 km² and 1286.1 km² respectively. A net increase of the urban land in the study area within four years was 372.06km², and the number was 192.6km² in the major cities; the increase of urban area in Hangzhou, Shaoxing and Ningbo was 43.7 km²; and that of the others was 145.6 km².It revealed that the small cities played a dominant role in the process of urbanization, but the bigger cities’ were relative weak. The urban land in Hangzhou and Shaoxing expanded obviously, and especially the expansion of Xiasha district in Hangzhou and Paojiang industrial park in Shaoxing were more remarkable. However, the smallest expansion occurred in Ningbo, especially in Yuyao urban district. In expansion direction, Hangzhou urban districts mainly developed toward northwest and north; while Ningbo urban district spread in all directions evenly.

Key words: remote sensing, urban expansion, the decision tree method, the index of shape, the region round Hangzhou Bay