摘要
选择东北地区典型内陆沼泽湿地——三江平原洪河自然保护区作为研究区,结合SAR的极化特性,分析了多时相ENVISAT ASAR不同极化下洪河湿地保护区不同地物植被类型的散射特性,利用长波L波段PAL-SAR数据对植被的可穿透性及水分的敏感性,结合与光学影像TM融合后进行神经元网络分类的方法,应用决策树方法进行了多波段、多时相SAR合成湿地植被识别试验。本文将两种方法相结合,分两步完整识别出沼泽、灌丛、岛状林、草甸、开阔水体及少量农田。
This paper selected typical northeast inland wetlands-Honghe nature reserve in Sanjiang Plain as the study area,and utilized the different vegetation backscatter characteristics of multi-temporal,multi-polarization of ASAR and the vegetation penetrability of PALSAR to identify marsh,bush,forest,grass,open water and farm land by combining the decision tree method and neural network method after fusing TM and PALSAR image.It can be seen that the radar system has ability of detecting hydrology,vegetation type and soil moisture information,connecting with the ＂wetness＂ information of wetland,and is a reliable data resource of wetland study.
出处
《遥感信息》
CSCD
2012年第2期15-19,共5页
Remote Sensing Information
基金
GIS和RS支持下的内陆平原淡水湿地水文生态模型研究(NSFC40871241)资助
