【作者】 陈洁丽； 申成磊； 刘永学； 李满春； 张栋；

【Author】 CHEN Jieli,SHEN Chenglei, LIU Yongxue, LI Manchun, ZHANG Dong (School of Geographic & Oceanographic Science, Nanjing University, Nanjing 210093)

【机构】 地理与海洋科学学院南京大学；

【摘要】 以无锡市作为研究区域,采用2000年Landsat ETM+影像数据,通过对居民地的遥感机理分析,利用植被指数、水体指数、城镇指数相结合的方法提取居民地信息。分析遥感影像的谱间结构特征,通过试验,当MNDWI>0.20时,可以很好地将水域信息提取出来;当NDBI-MNDWI>0.35时,可以较好提取出采石场;当NDBI-SAVI>0.23时,可以较好提取出居民地、水域和采石场。因此,在此基础上建立二值逻辑运算式:居民地=(·NOT·(MNDWI>0.2))·AND·(·NOT·(NDBI-MNDWI>0.35))·AND·(NDBI-SAVI>0.23),得到居民地遥感信息提取结果。此外,并用该方法在1995年的LandsatTM影像上进行了进一步的验证。研究结果表明:该方法可以将居民地信息提取出来,并且效果较好,是一种行之有效的方法。更多还原

【Abstract】 In this paper, taking Wuxi city as a case study area, selected Landsat ETM + image in 2000 as the data source. Based on the remote sensing mechanism analysis of the residential areas, vegetation index, water index, and the urban index are combined to extract the remote sensing information for residents. Analyzing of remote sensing image spectral signature features, when MNDWI>0.20, water can be extracted well; when NDBI-MNDWI>0.35, quarry areas can be extracted well; when NDBI-SAVI>0.23, residential areas, water and quarry areas can be extracted well. So using the two-valued logic expression : residential areas = (·NOT·(MNDWI>0.2))·AND·(·NOT·(NDBI-MNDWI>0.35)) ·AND·(NDBI-SAVI>0.23). It can get the result of residential areas. Further study is carried on the Landsat TM image in 1995 to verify the method. The results indicate that this model is effective with the high accuracy of the extraction.更多还原