Estimating wheat leaves chlorophyll content using hyperspectral technology and integrated inversion approach

Leaf chlorophyll content (LCC) is an important indicator for wheat growth monitor. In this paper, the integrated inversion approach was developed to estimate the wheat LCC. Based on the PROSAIL simulation data set, three hyperspectral vegetation indices (VIs), PRI, TCARI and TCARI/OSVAI, were comparative analyzed to screen out optimal VI for wheat LCC estimation. The integrated inversion models to estimate wheat LCC were built using the curve fit and least squares support vector regression (LS-SVR) algorithms as the modeling methods, respectively. Finally, using the observation dataset, the accuracy of LS-SVR model for wheat LCC inversion was verified. The result showed that TCARI/OSVAI was an optimal VI for LCC estimating, as it not only exhibit a good sensitivity to the change of LCC but also showed a least sensitivity to the change of LAI values among the three indices and therefore least affected by canopy density when used to estimate the wheat LCC; Comparative curve fitting algorithm, LS-SVR was a optimal algorithm for modeling, as indicated by higher R² (0.932 for LS-SVR model and 0.923 for curve fitting model) and lower RMSE (3.065 for LS-SVR model and 3.209 for curve fitting model); The R² of the fitting model between the estimated values and measured values reached 0.763, indicated the similarity between estimated and measured value was high, and it was feasible to obtain the wheat LCC accurately by using hyperspectral VIs and integrated inversion approach.