研究目的
To evaluate the feasibility of using vegetation indices, especially the red-edge reflectance based VIs derived from RapidEye data, for LAI estimation and to investigate the physical basis of a generic estimation model using a RE-based VI.
研究成果
The study concludes that RE-based VIs are more effective than VIS-based VIs for developing a generic LAI estimation model across different crops due to reduced sensitivity to canopy structure variations. This supports the use of satellite data with red-edge bands, such as from RapidEye and Sentinel-2, for crop monitoring.
研究不足
Limitations include uncertainties in field measurements of LAI and ALA due to the DHP method and CAN-EYE software processing, sensitivity of RE-based VIs to soil reflectance at low LAI, and the need to consider other factors like atmospheric conditions and sensor viewing geometry in future studies.
1：Experimental Design and Method Selection:
The study used a semi-empirical model based on modified Beer's law to characterize the relationship between LAI and VIs. The Markov Chain Monte Carlo (MCMC) sampling method was employed for parameter estimation, and sensitivity analysis was conducted using the PROSAIL model.
2：Sample Selection and Data Sources:
Field data were collected from spring wheat and canola crops in northern Ontario, Canada, over two years (2012 and 2013), including LAI, average leaf angle (ALA), and leaf chlorophyll content (Cab). RapidEye satellite images were used for remote sensing data.
3：List of Experimental Equipment and Materials:
Equipment included a digital hemispherical camera for LAI measurement, a CCM-200 chlorophyll meter for Cab measurement, and software such as CAN-EYE for data processing and PCI Geomatica for image preprocessing.
4：Experimental Procedures and Operational Workflow:
Field measurements were taken at specific plots, with LAI derived from hemispherical photos and Cab from chlorophyll meter readings. RapidEye images were atmospherically and geometrically corrected. VIs were calculated from reflectance data, and model parameters were estimated using MCMC.
5：Data Analysis Methods:
Statistical analysis included calculation of R2 and RMSE for model fits. Sensitivity analysis was performed using the Extended Fourier Amplitude Sensitivity Test (EFAST) with PROSAIL simulations.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
