基于DMSP/OLS数据的中国大尺度城镇用地信息提取  

Extraction of large-scale urban area information in China using DMSP/OLS nighttime light data

作　　者：王晓慧[1,2] 肖鹏峰[1,2] 冯学智[1,2] 李海星[1,2] 

机构地区：[1]江苏省地理信息技术重点实验室,南京大学,南京210093 [2]南京大学地理信息科学系,南京210093

出　　处：《国土资源遥感》2013年第3期159-164,共6页Remote Sensing for Land & Resources

基　　金：全球变化研究国家重大科学研究计划课题(编号:2011CB952001)

摘　　要：DMSP/OLS夜间灯光数据已在城镇用地信息提取中得到较多应用,其中,较高分辨率数据比较法是一种较为客观的方法,但现有研究主要通过逐个城市确定灯光数据的阈值提取城镇用地信息,这对于大尺度的城镇研究所需成本较大、处理效率较低。为此,提出了一种基于分区的较高分辨率数据比较法,将中国按经济发展程度划分为7大区,分别对各区提取城镇用地信息,并对提取结果进行真实城镇像元判定,得到1992,1995,2000年中国的城镇用地信息。结果表明,采用分区的方法不仅提高了较高分辨率数据比较法的提取效率,反映了特定区域的城镇整体状况,而且在总体区域上保证了提取精度。Nighttime light data have been widely used for mapping urban land.Previous application of the TM assisted thresholding method mainly determined threshold of the nighttime light data in each city,which is not suitable for the study at a large scale due to its low processing efficiency and high data acquisition cost.In this paper,based on the TM-assisted thresholding method,the authors put forward the partition idea to extract the urban area using the DMSP / OLS nighttime light data.China was divided into 7 major regions based on economic partition and then TM-assisted thresholding method was employed to extract urban land areas.Finally,applying real urban land use pixels judgment,the authors obtained China’s urban land information of the years 1992,1995 and 2000.A comparison with the previous method shows that the performance of the method based on partition idea is more efficient and can ensure overall accuracy.

关 键 词：DMSP OLS 较高分辨率数据比较法 城镇用地 分区 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]