基于LSMM和改进的FCM提取城市植被覆盖度——以北京市海淀区为例.pdfVIP

生态学报却始，础的:1018-1但4 Acta ECOI嚼皿Siniea 基于 LSMM 和改进的 FCM 提取城市植被覆盖度 一一以北京市海淀区为例 胡妹持，胡德、勇*，赵文吉 (三维信息获取与应用教育部重点实验室，北京 剧沼48;资源环境与地理信怠系统北京市重点实验室 首都师范大学资源环境与旅援学院，北京 I仪胆的 摘要:植被是城市生态系统的重要组成部分，及时在取植被覆盖信息对壤市生惠环境监拥具有重要意义。和;用中分辨率 Landsat TM 遥感数据，采用线性光谱分解模型(LSMM)开展城市植被覆盖度提取z 同时，通过改进训练样本选择方式，在最小噪 声变换(MNF) 、像孟纯净指数分析(PPI) 、 N 维可视化分析基础上得到端元样本，再运用模糊c-均值(FCM)获取植被覆盖度;最 后以离分辨率SP0T5遥感数据对两种方式的提取结果进行精度检验。结果显示，基于LSMM 和改造的FCM 提取的城市植被覆 盖度与检验数据相关系数分慰为0.8252 和0.9381 ，后者可以较好地处理其他要素的非线性影嘀，自而具有较高精度。 关键词:植被覆盖度;混合像元分解:线性光谱握合模型;改造的 FCM Extract urban vege臼tion coverage based on LSMM and improved FCM: a case study in Haidian District HU Shujing , HU Deyong* , ZHAO Wenji Key Labora郎ry of3D Ii吨阳阳tionAcq时ωn and Application ofMini蹄..， of Education , Be价19 l(削4窑 ， China Key 却bora如ry of Resour础 Em翩翩翩and G1S of Be.展现gMun崎al ， Beijil容:创那48 ， ChÏlla College of ReoUJ;出 Erwironment and Tourism , C甲tal Norrrwl UnweT$i妙， Beijir啄 1仅览48 ， China Abstract: Vegetation is an impo民缸1t part of 山乞ane∞system; therefore timely access to vegetation cover:艳e information is of 伊at significance for monitorin辜 的四 ecological environment. Linear spectral mixture model (LSMM) was carried out for urban vegetation CQ亨erage extractÍon using medium-r;恼。，lution Landsat TM remote sensing data. Meanwhile , the fuzzy c- me四s (FCM) method was chosen to e:双r配t vegetatÍon cover唔e hy improving the traini吨sample selectÍon method to obtaìn the end-member sample hased on minimum noise transform (MNF) , pixel purity index 部18lysis (PPI) , and N-dimensional visualization analysis. Finally , high-纪回，lution SPσ巴 remote sensing data extracted in two ways were 出ed to carry out the accuracy test for veget础。aωIver鸣e. 它le results showed that the correlation coe伍cients 如tween the inspection data and LSMM-hased and improved FCM-based data were