Estimating conifer forest LAI with HyMap data using a reflectance model and artificial neural nets

The potential of canopy reflectance modelling to retrieve structural variables in managed Norway spruce stands was investigated using the invertible forest reflectance model INFORM. INFORM was derived by coupling FLIM, SAIL and LIBERTY models, and was inverted with hyperspectral airborne HyMap data using a neural network approach. A relatively simple three layer feed-forward backpropagation neural network with two input neurons, one neuron in the hidden layer and three output neurons was employed. Leaf area index (LAI) field measurements from 39 forest stands were used to validate the LAI estimates produced from HyMap reflectances. Using two HyMap wavebands at 837 nm and 1148 nm the obtained accuracy of LAI amounts to an rmse of 0.58 (relative rmse 18 % of mean). In contrast to approaches based on empirical relations between a spectral vegetation index and the biophysical variable of interest, the inversion approach is applicable to various sensor types and site conditions.