Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
8866523 | Remote Sensing of Environment | 2018 | 7 Pages |
Abstract
Chlorophyll (Chl) is an important indicator of photosynthetic capacity and stress of vegetation. Remote sensing provides fast and nondestructive methods for estimating leaf Chl content based on its optical characteristics in visible and near-infrared spectrum. Multispectral lidar (MSL) systems have been developed to combine spectral and spatial detection abilities. Statistical relationships of plant biochemical constituents can be established through MSL measurements. However, empirical models cannot be readily extended to independent datasets. Simultaneously, the few spectral bands of MSL limit the use of a physical model. Hence, the development of hyperspectral lidar (HSL) systems offers a wider range of spectrum. This study investigated the possibility of adopting an HSL system with 32 channels covering 539-910â¯nm to estimate foliar Chl through a physical model. This study aimed to (1) Determine whether reflectance at the 32 channels is sufficient to retrieve Chl content through PROSPECT model inversion and (2) Considering the difference between passively and actively measured reflectance, investigate whether HSL measurements can be applied into PROSPECT model inversion for leaf biochemical constituents. Three kinds of datasets were used: a synthetic dataset simulated by running the PROSPECT model in forward mode, a public dataset ANGERS taking the channels of the HSL system, and an experimental dataset of paddy rice measured by the HSL system. Results showed HSL measurements can be directly used to retrieve leaf Chl content through PROSPECT-4 model inversion (R2â¯=â¯0.55). These measurements also exhibit higher accuracy than that of support vector regression (threefold cross validation; 100 repetitions: median R2â¯=â¯0.47). This validation work provides basis in the determination of vegetation physiological status directly from HSL measurements through model inversion with the PROSPECT model.
Related Topics
Physical Sciences and Engineering
Earth and Planetary Sciences
Computers in Earth Sciences
Authors
Jia Sun, Shuo Shi, Jian Yang, Biwu Chen, Wei Gong, Lin Du, Feiyue Mao, Shalei Song,