摘要
地物光谱数据是多因素共同作用的结果,可能会出现"同物异谱"或"异物同谱"现象。以科尔沁沙地典型沙丘—草甸相间区域为研究区,以野外实测沙丘植被光谱数据为基础,利用最小距离法、光谱相似角度、光谱相关系数和光谱信息散度4种光谱匹配算法,对去噪后植物光谱特征数据进行匹配分析,发现最小距离法可以区分出本区沙丘典型植被。在此基础上,对光谱导数进行匹配分析,发现光谱相似角度和光谱相关系数识别效果增强明显,二阶导数光谱特征匹配度甚至出现了多个负值。最后选取TM1、TM2、TM3和TM4波段植被光谱数据做同样处理,证实TM2、TM4波段能够提高识别精度;同时,基于统计学基础分析单波段基础信息,发现单波段所包含信息量依次为TM1>TM4>TM2>TM3,综合考虑TM1波段内各植被光谱反射率特征曲线重叠严重和单波段植被光谱特征匹配度结果,认为TM4波段为该地区植被物种识别的首选波段。
Spectral data are the joint affecting results of many factors, and there may be the ＂same objects with different spectra＂ or ＂different objects with the same spectra＂. In this study, four common spectral matching algorithms, i.e. the Minimum Distance （ MD）, Spectral Similarity Angle （ SSA）, Spectral Correlation Coefficient （SCC） and Spectral Information Divergence （SID）, were used to carry out the matching analysis on the denoised spectral data of plants growing on sand dune in the Horqin Sandy Land. It was found that the minimum distance method could be used to distinguish the typical plant species on sand dune in the study area. On which a matching analysis of spectral derivative was carried out, it was found that the distinguishing effects of spectral similar angle and spectral correlation coefficient were significantly increased, and even a negative value of the matching degree of second derivative spectrum occurred. The band TM1, TM2, TM3 and TM4 of vegetation spectral data were selected to process, and it was verified that the band TM2 and TM4 could be used to improve the distinguishing accuracy. After analyzing the basic information of single band with the statistically based analysis, it was found that the amount of information contained in a single band was in an order of TM1 〉 TM4 〉 TM2 〉 TM3, and it was considered that the band TM4 should be the first choice to distinguish the plant species in the study area based on a comprehensive consideration of the highly overlapped spectral reflectance curve and the spectral matching results of single band of plant species.
出处
《干旱区研究》
CSCD
北大核心
2014年第4期750-755,共6页
Arid Zone Research
基金
国家科技部国际合作项目(2010DFA71460)
国家自然科学基金重点项目(51139002)
国家自然科学基金(51069005)
国家自然科学基金(51269014)
内蒙古自治区自然科学基金(2011BS0607)
内蒙古农业大学寒旱区水资源利用创新团队项目(NDTD2010-6)
关键词
植被光谱特征
光谱匹配度
光谱导数
TM波段
科尔沁沙地
spectrum characteristics of vegetation
spectral matching analysis
spectra derivative
TM band
Horqin Sandy Land
