›› 2012, Vol. 20 ›› Issue (6): 1004-1010.DOI: 10.11733/j.issn.1007-0435.2012.06.004

• 研究论文 • 上一篇    下一篇

基于反射率及导数的草原植被冠层光合有效吸收分量高光谱反演

刘爱军1,2,3, 王保林2,3, 黄平平4, 卢欣石1   

  1. 1. 北京林业大学, 北京 100083;
    2. 内蒙古草原勘察规划院, 内蒙古 呼和浩特 010051;
    3. 内蒙古民族大学, 内蒙古 通辽 028000;
    4. 内蒙古工业大学, 内蒙古 呼和浩特 010051
  • 收稿日期:2012-06-24 修回日期:2012-10-19 出版日期:2012-12-15 发布日期:2012-12-28
  • 通讯作者: 卢欣石
  • 作者简介:刘爱军(1965-),女,内蒙古赤峰人,研究员,研究方向为草原生态、遥感技术与应用,E-mail:liuaj_81@163.com;
  • 基金资助:
    农业行业科研专项(200903060);林业行业科研专项(201204202);农业科技成果转化资金项目(2011A4010001)资助

Research on fPAR Estimation of Grassland with Hyperspectral Data Based on Reflectivity and Derivative

LIU Ai-jun1,2,3, WANG Bao-lin2,3, HUANG Ping-ping4, LU Xin-shi1   

  1. 1. Beijing Forestry University, Beijing 100083, China;
    2. Inner Mongolia Institute of Grassland Survey and Design, Hohhot, Inner Mongolia 010051, China;
    3. Inner Mongolia University for Nationalities, Tongliao, Inner Mongolia 028000, China;
    4. Inner Mongolia University of Technology, Hohhot, Inner Mongolia 010051, China
  • Received:2012-06-24 Revised:2012-10-19 Online:2012-12-15 Published:2012-12-28

摘要: 利用ASD便携式野外光谱仪和光量子仪实测了6种草原植被类型关键生育期的反射光谱数据和光合有效辐射值,利用可见光波段处反射率及近红外波段区处一阶导数分别与fPAR建立逐步回归方程,同时,将各波段反射率与各波段导数光谱建立逐步回归方程。结果表明:典型草原光合有效辐射分量与可见光反射率相关性好于近红外波段反射率,其中在405和470 nm波段相关性最好;fPAR与一阶导数相关关系在855和965 nm波段处较强。fPAR与405和470 nm反射率以及965 nm一阶导数的多波段逐步回归分析结果取得了较单波段和NDVI最优的估算效果,R2达0.939。利用高光谱数据进行fPAR估算时,需要综合考虑可见光和近红外波段信息,同时也要充分考虑反射率与反射率导数的方法。水分强吸收的光谱波段具有提高fPAR估算精度的潜力。

关键词: 草原植被, 冠层, 光合有效吸收分量, 高光谱数据

Abstract: Hyperspectral reflectance and fraction of absorbed photosynthetically active radiation (fPAR) of six grassland types were measured by ASD portable spectrometer and light quantum respectively in key growth stages. Relativity between hyperspectral reflectance and first-order derivative were analyzed in order to estimate fPAR and establish the best estimation model. Results showed that the maximum correlation coefficients between reflection values and fPAR occurred at 405 nm and 470 nm wavelength, respectively. Correlation between the first derivative and fPAR was better at 855nm and 965nm, respectively. Stepwise regression equation between fPAR and reflectivity values (in 405 nm, 470 nm wavelength) as well as first-order derivative in 965 nm wavelength was established. These results showed that stepwise regression of multiband, which R2 was 0.939, obtained more optimal estimation effect than that of single band and NDVI. High spectral resolution data could distinguish sensitive bands for vegetation spectrum, which was essential to select appropriate band and first-order derivative for calculating fPAR. In addition, the spectral band of strong water absorption was a potential to improve estimation accuracy of fPAR.

Key words: Grassland, Canopy, fPAR, Hyperspectral data

中图分类号: