APPLICATION OF AN ARTIFICIAL NEURAL-NETWORK IN CANOPY SCATTERING INVERSION

Owing to their recent success in other inversion tasks, application of an artificial neural network to the development of an inversion algor ithm for radar scattering from vegetation canopies is considered. Beca use canopy scattering models are complicated functions of the desired biophysical parameters (vegetation biomass, leaf area index, soil mois ture content, etc.), the development of an effective inversion algorit hm is not a straightforward task. The Michigan Microwave Canopy Scatte ring (MIMICS) model, which has shown remarkable success in predicting the radar response to vegetation canopies, was used, as were measured polarimetric backscatter values. Hence, the radiative transfer simulat ion code, MIMICS, was used to produce some of the training data. The i nputs to the neural network were the expected polarimetric backscatter values from specific canopies, while the outputs were the desired par ameters, such as tree heights, crown thickness, leaf density, etc. Two special cases were examined: (1) inversion of MIMICS given modelled a spen stands of different ages; (2) inversion of measured data from the Duke forest loblolly pine stands. The MIMICS inversion shows that neu ral networks are capable of accurately inverting some of the parameter s of such a complicated model. The implication is that once MIMICS is made to model the radar data for a specific application, then inversio n of the radar data may be accomplished. The measured data inversion s hows that, even without a model such as MIMICS, one may train a neural network to invert several parameters of interest. However, this depen ds on accurate and complete surveys of the ground truth data to be use ful.