中文摘要：

不透水面积的变化是城市化进程的一个重要表征。本文选取高反射率、低反射率、植被和土壤四种地类端元,使用完全受限的线性光谱分解算法直接从Landsat TM5影像中提取了1995年和2007年北京市平原区的不透水面积比例及其空间分布信息。利用2006-2007年的大比例电子地图验证了2007年的不透水面积分布信息提取结果,总的均方差达到17.3%。通过对两个时间的不透水面积比例进行对比,分析了1995年到2007年北京城市化进程的主要范围和规模,结果表明顺义、通州、房山、大兴、朝阳和昌平等区是城市化进程比较显著的地区。本文直接利用遥感数据获取较高精度的不透水面积信息,结果可为分析城市化进程、城市水文和能量循环等提供重要基础性数据。

英文摘要：

Change in impervious ratio of ground surface is an important index of urbanization. This paper presents a direct estimation of impervious surface distribution in the plain area of Beijing from landsat thematic mapper （ET） data of 1995 and 2007 with a fully constrained linear spectral mixture model of four endmembers, i.e. high albedo, low albedo, vegetation, and soil. The estimation accuracy of this ratio in 2007 was evaluated by using a high-precision geographic vector map of 2006 ~ 2007, and an overall RMS error of 17.3% was obtained. Comparison of the impervious surface distributions in 1995 and 2007 shows the main regions of urbanization in Beijing and their scales. Results show that in this period the urbanization were prominent in Shunyi, Tongzhou, Fangshan, Daxing, Chaoyang and Changping. This direct estimation method from Landsat TM images provides important information to the studies of urbanization, urban hydrological and energy cycle.