Back to Search Start Over

Estimation of canopy carotenoid content of winter wheat using multi-angle hyperspectral data.

Authors :
Kong, Weiping
Huang, Wenjiang
Liu, Jiangui
Chen, Pengfei
Qin, Qiming
Ye, Huichun
Peng, Dailiang
Dong, Yingying
Mortimer, A. Hugh
Source :
Advances in Space Research. Nov2017, Vol. 60 Issue 9, p1988-2000. 13p.
Publication Year :
2017

Abstract

Precise estimation of carotenoid (Car) content in crops, using remote sensing data, could be helpful for agricultural resources management. Conventional methods for Car content estimation were mostly based on reflectance data acquired from nadir direction. However, reflectance acquired at this direction is highly influenced by canopy structure and soil background reflectance. Off-nadir observation is less impacted, and multi-angle viewing data are proven to contain additional information rarely exploited for crop Car content estimation. The objective of this study was to explore the potential of multi-angle observation data for winter wheat canopy Car content estimation. Canopy spectral reflectance was measured from nadir as well as from a series of off-nadir directions during different growing stages of winter wheat, with concurrent canopy Car content measurements. Correlation analyses were performed between Car content and the original and continuum removed spectral reflectance. Spectral features and previously published indices were derived from data obtained at different viewing angles and were tested for Car content estimation. Results showed that spectral features and indices obtained from backscattering directions between 20° and 40° view zenith angle had a stronger correlation with Car content than that from the nadir direction, and the strongest correlation was observed from about 30° backscattering direction. Spectral absorption depth at 500 nm derived from spectral data obtained from 30° backscattering direction was found to reduce the difference induced by plant cultivars greatly. It was the most suitable for winter wheat canopy Car estimation, with a coefficient of determination 0.79 and a root mean square error of 19.03 mg/m 2 . This work indicates the importance of taking viewing geometry effect into account when using spectral features/indices and provides new insight in the application of multi-angle remote sensing for the estimation of crop physiology. [ABSTRACT FROM AUTHOR]

Details

Language :
English
ISSN :
02731177
Volume :
60
Issue :
9
Database :
Academic Search Index
Journal :
Advances in Space Research
Publication Type :
Academic Journal
Accession number :
125419873
Full Text :
https://doi.org/10.1016/j.asr.2017.07.015