香格里拉4种典型针叶树种高光谱特征分析及判别
|
摘要
以香格里拉市云南松、高山松、云杉和冷杉4种典型针叶树种为研究对象,运用ASD Field Spec 3地物光谱仪测定野外叶片光谱,并对原始光谱进行微分变换处理,再采用Fisher判别分析方法对4种针叶树种最佳波段窗口进行分析.结果表明:Fisher判别分析能有效判别4种典型针叶树种原始光谱、一阶微分光谱和二阶微分光谱差异显著的波段,主要位于近红外波段,最佳波段窗口分别为980～989、415～424、960～969 nm;原始光谱的二阶微分处理更能有效判别4种针叶树种,Fisher总判别精度高达98.8%;根据4种典型针叶树空间分布特征,将其分为云南松、高山松和云冷杉两组,云南松、高山松的最佳波段窗口为870～879、1 020～1 029、530～539 nm,云杉、冷杉的最佳波段窗口为540～549、520～529、1 150～1 159 nm.本研究结果可为中大尺度机载、星载高光谱遥感树种精细分类提供指导.
The original spectrum of 4 typical conifer species in Shangri-la City, namely Pinus yunnanensis, Pinus densata, Picea asperata and Abies fabri, were measured by ASD Field Spec 3 spectrometer in November. Then, first-order differential and second-order differential were applied to the original spectrum data to identify characteristic bands for each species. Lastly, Fisher discriminant analysis was used to select the optimal band window among characteristic bands that effectively distinguish species one from another. The results show that near-infrared region was the waveband that showed significant differences in the original spectrum, first-order differential spectrum and second-order differential spectrum which can be effectively distinguished by Fisher discriminant analysis. The characteristic bands detected were 980-989, 415-424, 960-969 nm. Second-order differential turned out to be the most effective data processing method with Fisher′s total discriminant accuracy being 98.8%. Four species were regrouped into P.yunnanensis, P.densata and P.asperata, A.fabri according to their spatial distribution characteristics. The optimal band windows for distinguishing P.yunnanensis from P.densata were 870-879, 1 020-1 029, 530-539 nm, while the optimal band windows for distinguishing P.asperata from A.fabri were 540-549, 520-529, 1 150-1 159 nm.
The original spectrum of 4 typical conifer species in Shangri-la City, namely Pinus yunnanensis, Pinus densata, Picea asperata and Abies fabri, were measured by ASD Field Spec 3 spectrometer in November. Then, first-order differential and second-order differential were applied to the original spectrum data to identify characteristic bands for each species. Lastly, Fisher discriminant analysis was used to select the optimal band window among characteristic bands that effectively distinguish species one from another. The results show that near-infrared region was the waveband that showed significant differences in the original spectrum, first-order differential spectrum and second-order differential spectrum which can be effectively distinguished by Fisher discriminant analysis. The characteristic bands detected were 980-989, 415-424, 960-969 nm. Second-order differential turned out to be the most effective data processing method with Fisher′s total discriminant accuracy being 98.8%. Four species were regrouped into P.yunnanensis, P.densata and P.asperata, A.fabri according to their spatial distribution characteristics. The optimal band windows for distinguishing P.yunnanensis from P.densata were 870-879, 1 020-1 029, 530-539 nm, while the optimal band windows for distinguishing P.asperata from A.fabri were 540-549, 520-529, 1 150-1 159 nm.
引文
[1] 王志辉,丁丽霞.基于叶片高光谱特性分析的树种识别[J].光谱学与光谱分析,2010,30(7):1 826-1 830.
[2] 洪娇,舒清态.滇西北高寒山区云冷杉高光谱差异性研究[J].西北林学院学报,2017,32(2):252-255.
[3] 褚西鹏,葛宏立,陈柯萍.基于小波变换的叶片高光谱数据的树种分类[J].光谱实验室,2012,29(5):2 795-2 799.
[4] 丁丽霞,王志辉,葛宏立.基于包络线法的不同树种叶片高光谱特征分析[J].浙江林学院学报,2010,27(6):809-814.
[5] 刘秀英,臧卓,孙华,等.基于高光谱数据的杉木和马尾松识别研究[J].中南林业科技大学学报,2011,31(11):31-34.
[6] PROSPERE K, MCLAREN K, WILSON B. Plant species discrimination in a tropical wetland using in situ hyperspectral data[J]. Remote Sensing, 2014,6(9):8 494-8 523.
[7] MENG R, DENNISON P E. Spectroscopic analysis of green, desiccated and dead tamarisk canopies[J]. Photogrammetric Engineering and Remote Sensing, 2015,81(3):199-207.
[8] 李圣娇.基于高光谱遥感的不同针叶林类型的光谱分析与识别研究[D].昆明:西南林业大学,2016.
[9] 许师凯,王基,刘树勇,等.基于Savitzky-Golay算法的混沌平滑去噪[J].环境工程,2014,32:105-109.
[10] 闫晓勇.基于冠层光谱的南疆盆地主栽果树树种识别有效波段选择研究[D].乌鲁木齐:新疆农业大学,2014.
[11] 吴见,彭建,王孟和,等.几种常见树种叶片光谱秋季变化特征分析[J].光谱学与光谱分析,2017,37(4):1 226-1 232.
[12] 徐光彩,庞勇,李增元,等.小兴安岭主要树种冠层光谱季相变化研究[J].光谱学与光谱分,2013,33(12):3 304-3 309.
[13] 李伟涛,彭道黎,张艳,等.琅琊山区主要树种冠层光谱年际变化研究[J].光谱学与光谱分析,2015,35(8):2 247-2 233.
[14] 江春梅,陈文惠,黄传印.基于实测冠层光谱数据的三明市13种树种识别研究[J].亚热带资源与环境学报,2016,11(2):58-67.
[15] 李静萍,谢邦昌.多元统计分析方法与应用[M].北京:中国人民大学出版社,2008.
[16] 齐浩,王振锡,岳俊,等.基于叶片光谱特征的南疆盆地主栽果树树种遥感判别[J].浙江农业学报,2015,27(12):2 141-2 146.
[17] JONES T G, SHARMA T, SHARMA T. Employing ground-based spectroscopy for tree-species differentiation in the Gulf Islands National Park Reserve[J]. International Journal of Remote Sensing, 2010,31(4):1 121-1 127.
[18] CROSS M D, SCAMBOS T A, PACIFACI F, et al. Validating the use of metre-scale multi-spectral satellite image data for identifying tropical forest tree species[J]. International Journal of Remote Sensing, 2018,39(11):3 723-3 752.
[2] 洪娇,舒清态.滇西北高寒山区云冷杉高光谱差异性研究[J].西北林学院学报,2017,32(2):252-255.
[3] 褚西鹏,葛宏立,陈柯萍.基于小波变换的叶片高光谱数据的树种分类[J].光谱实验室,2012,29(5):2 795-2 799.
[4] 丁丽霞,王志辉,葛宏立.基于包络线法的不同树种叶片高光谱特征分析[J].浙江林学院学报,2010,27(6):809-814.
[5] 刘秀英,臧卓,孙华,等.基于高光谱数据的杉木和马尾松识别研究[J].中南林业科技大学学报,2011,31(11):31-34.
[6] PROSPERE K, MCLAREN K, WILSON B. Plant species discrimination in a tropical wetland using in situ hyperspectral data[J]. Remote Sensing, 2014,6(9):8 494-8 523.
[7] MENG R, DENNISON P E. Spectroscopic analysis of green, desiccated and dead tamarisk canopies[J]. Photogrammetric Engineering and Remote Sensing, 2015,81(3):199-207.
[8] 李圣娇.基于高光谱遥感的不同针叶林类型的光谱分析与识别研究[D].昆明:西南林业大学,2016.
[9] 许师凯,王基,刘树勇,等.基于Savitzky-Golay算法的混沌平滑去噪[J].环境工程,2014,32:105-109.
[10] 闫晓勇.基于冠层光谱的南疆盆地主栽果树树种识别有效波段选择研究[D].乌鲁木齐:新疆农业大学,2014.
[11] 吴见,彭建,王孟和,等.几种常见树种叶片光谱秋季变化特征分析[J].光谱学与光谱分析,2017,37(4):1 226-1 232.
[12] 徐光彩,庞勇,李增元,等.小兴安岭主要树种冠层光谱季相变化研究[J].光谱学与光谱分,2013,33(12):3 304-3 309.
[13] 李伟涛,彭道黎,张艳,等.琅琊山区主要树种冠层光谱年际变化研究[J].光谱学与光谱分析,2015,35(8):2 247-2 233.
[14] 江春梅,陈文惠,黄传印.基于实测冠层光谱数据的三明市13种树种识别研究[J].亚热带资源与环境学报,2016,11(2):58-67.
[15] 李静萍,谢邦昌.多元统计分析方法与应用[M].北京:中国人民大学出版社,2008.
[16] 齐浩,王振锡,岳俊,等.基于叶片光谱特征的南疆盆地主栽果树树种遥感判别[J].浙江农业学报,2015,27(12):2 141-2 146.
[17] JONES T G, SHARMA T, SHARMA T. Employing ground-based spectroscopy for tree-species differentiation in the Gulf Islands National Park Reserve[J]. International Journal of Remote Sensing, 2010,31(4):1 121-1 127.
[18] CROSS M D, SCAMBOS T A, PACIFACI F, et al. Validating the use of metre-scale multi-spectral satellite image data for identifying tropical forest tree species[J]. International Journal of Remote Sensing, 2018,39(11):3 723-3 752.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |