基于高光谱数据的鄱阳湖湿地典型植被识别分析  

Discriminant analysis of typical vegetation species in Poyang Lake wetland based on hyperspectral data

作　　者：陈彦兵[1] 况润元[1] 曾帅 CHEN Yanbing;KUANG Runyuan;ZENG Shuai(School of Architectural and Surveying&Mapping Engineering,Jiangxi University of Science and Technology,Ganzhou 341000,China)

机构地区：[1]江西理工大学建筑与测绘工程学院,江西赣州341000

出　　处：《人民长江》2018年第20期19-23,共5页Yangtze River

基　　金：国家自然科学基金项目(41101322);江西省自然科学基金项目(20114BAB213022)

摘　　要：利用高光谱数据对湿地植被进行分类历来是植被遥感研究的重点和难点之一。以鄱阳湖为研究区,测取了5种典型湿地植被的高光谱数据,利用光谱微分法对原始光谱数据进行处理,分析不同植被原始光谱、一阶微分和二阶微分光谱曲线图;使用欧式距离法选择差异较大的波段以鉴别不同植被;最后利用马氏距离法检验所选择波段识别不同植被的效果。结果表明:3种光谱曲线所提取的差异性较大,波段虽有所差异,但多位于近红外波段,马氏距离在这些波段上能有效对不同典型湿地植被进行识别。研究结果可为湿地植被分类识别奠定基础,同时可为湖泊湿地植被以及湖泊生态环境的保护决策提供科学依据。Classification of the wetland vegetation by hyperspectral data is one of the most difficult and important aspect in vegetation remote sensing.By measuring hyperspectral data of 5 typical vegetations in Poyang Lake,in this paper,derivative reflectance method was used to deal with the original spectral data.We analyze and compare the original spectrum curves,the first and second derivative reflectance of the five wetland vegetables.By analyzing the curves,we select the obvious different bands to identify plants by Euclidean distance,and then use Mahalanobis distance to judge the identification effect of the selected bands.The results show that the identification differences of the 3 spectrum curves are distinguished;the bands concentrate near the infrared band,where Mahalanobis Distance can effectively identify the typical wetland vegetation.The results not only lay a foundation for wetland vegetation classification,but also provide scientific basis for the protection of lake wetland vegetation and ecological environment.

关 键 词：湿地植被识别 高光谱数据 马氏距离 鄱阳湖 

分 类 号：X17[环境科学与工程—环境科学]