Inverting the PROSAIL canopy reflectance model using neural nets trained on streamlined databases

Clement Atzberger
2010 Journal of Spectral Imaging  
all rights reserved JSI By providing both spatial and temporal information, remote sensing may function as a cost-effective source of data for precision agriculture with potentially positive benefits for both the farmer and the environment. 1,2 amongst other bio-physical variables, the leaf area index (LAI) is an important indicator of actual crop growth, because it is related to ground coverage and photosynthetically active radiation (Par) interception efficiency. 3,4 remotely sensed maps of
more » ... e LAI are therefore considered helpful for decision making and yield predictions. 5,6 To assess crop productivity at regional to global scales, mapped temporal LAI profiles were, for example, assimilated into mechanistic crop growth models 7 or integrated into Par interception formulae. 8 The canopy LAI can be spectrally estimated by means of empirical-statistical methods 9,10 or through inversion of physically-based radiative transfer models. 11-15 Most empirical-statistical methods rely on vegetation indices, either formulated using classical red and near infrared (nIr) wavebands 16 or optimised through band selection in the hyperspectral feature space. [17] [18] [19] Still other studies use features in the red-edge to estimate canopy bio-physical variables 20,21 or establish multivariate regression approaches. 22-24 The empirical-statistical approaches have the advantage that they are relatively simple and easy to use, while reasonably accurate results can be obtained. The main drawbacks of these The widely used PROSAIL radiative transfer model was coupled with a simple soil reflectance parameterisation to estimate the leaf area index (LAI) of winter wheat (Triticum aestivum) from ground-based spectrometer data. To avoid time-consuming numerical optimisations, a neural net (NN) was used for model inversion. The NN was trained on 3000 spectral patterns generated by the reflectance model. The training database was previously streamlined to provide good approximation of the response surface while keeping the net compact. Streamlining was achieved by retaining only those synthetic spectra that belong both to the simulated and actual measurement spaces. The estimated LAI (n obs = 15) compared well with completely independent reference measurements taken four times during the 2000 growing season in four commercial winter wheat fields (1.8 ≤ LAI ≤ 8.1). The coefficient of determination (R 2 ) between measured and estimated LAI was 0.87 with a root mean squared error (RMSE) of 0.89 (m 2 m -2 ). Even for LAIs exceeding 3-4, saturation effects were low. Three measurement dates yielded RMSE lower than 0.8. Only during stem elongation did RMSE exceed 1. Higher errors for this time period were attributed to abrupt changes in the canopy structure (i.e. average leaf angle) not taken into account. Compared to the normalised difference vegetation index (NDVI), the inversion of PROSAIL using hyperspectral reflectances performed well, with errors reduced by more than 50% as compared to the NDVI model (RMSE: 1.91 m 2 m -2 ).
doi:10.1255/jsi.2010.a2 fatcat:gcpfb64ny5gkjkc4opohooh2j4