基于包络线去除法的贵州省常见树种高光谱特征分析

Analysis of Hyperspectral Characteristics of Common Tree Species in Guizhou Province Based on Envelope Removalmethod

为区别喀斯特高原森林植被类型,利用高光谱技术对森林植被开展光谱特征分析,为喀斯特高原贵州省森林植被的遥感探测与精准识别提供理论与技术支持。本研究采用ASDFieldSpec4便携式地物光谱仪对贵州省常见的华山松(Pinus armandiiFranch)、麻栎(Quercus acutissimaCarruth)、马尾松(Pinus massonianaLamb)、杉木(Cunninghamia lanceolata)、水杉(Metasequoia glyptostroboides)、香樟(Cinnamomum camphora)、云南松(Pinus yunnanensis)、柏木(Cupressus funebris Endl.)8种树种的光谱数据进行采集。在对原始光谱数据进行异常值处理的基础上,进行包络线去除,比较原始光谱和包络线去除图,选择差异较大的波段用于识别不同树种。结果表明:原始光谱中11个特征波段差异都较为明显,去包络线中1439~1448 nm、1779~1788 nm、1958~1967 nm和2307~2316 nm波段范围内差异较为明显,说明包络线去除法能有效缩小树种特征波段范围,达到运用较少波段识别树种的目的。用欧氏距离法对选取的特征波段进行验证,发现运用原始光谱曲线与包络线去除法选取的特征波段皆能有效区分树种,且包络线去除法的效果更好。

In order to distinguish forest vegetation types in Karst Plateau,spectral characteristics of forest vegetation were analyzed by hyperspectral technology,which provided theoretical and technical support for remote sensing detection and accurate identification of forest vegetation in Guizhou Province.In this study,ASD FeldSpec4 portable ground object spectrometer was used to collect the spectral data of eight common tree species in Guizhou Province,including Pinus armandii Franch,Quercus acutissima Carruth,Pinus massonianaLamb,Cunninghamia lanceolate,Metasequoia glyptostroboides,Cinnamomum camphora,Pinus yunnanensis and Cupressus funebris Endl.On the basis of processing the outliers of the original spectral data,we remove the envelope,compare the original spectrum and the envelope removal chart,select different bands to identify different tree species.The results show that the difference of 11 characteristic bands in the original spectrum is obvious,and the difference of 1439~1448 nm,1779~1788 nm,1958~1967 nm and 2307~2316 nm in the de enveloping line is obvious.The results show that the envelope method can effectively reduce the range of tree species characteristic band,and achieve the purpose of using less band to identify tree species.After the Euclidean distance method is used to verify the selected characteristic band,it is found that the characteristic band selected by the original spectral curve and envelope division method can effectively distinguish tree species,and the effect of envelope division method is better.