Retrieval of vegetation understory information fusing hyperspectral and panchromatic airborne data

The knowledge of the characteristics of the vegetation cover is of great interest in climate change process understanding due to its important role in controlling water and carbon cycles. The properties of vegetated surfaces are usually estimated from remote observations through semi-empirical regression models or using radiative transfer models, which simulate the interactions of solar radiation with the vegetated medium. In real domain the spectral responses measured by the sensor in forested area are strongly influenced by the different natural understory conditions that limit the possibility of applying both retrieval methods to predict overstory vegetation parameters. Understory information is therefore needed for trees parameter estimation; moreover from a biodiversity point of view and forest management perspective understory represents a critical component of forest ecosystem that needs a better characterization. An experiment was conducted using multisensor hyperspectral and panchromatic data to study the typology and status of different vegetated understory under a sparse poplar plantation. The salient aspect of the method is the integration of spectral (multisensor) domain fusion and spatial domain fusion techniques within a Multi-Layer Perceptron model. Ground data and airborne hyperspectral imagery were collected during DARFEM experiment, EU-funded HySens project (DLR-Germany). The achieved results show that this methods is able to solve “operatively” the problem of a volumetric mixture typical of natural forest ecosystems identifying the different surfaces present under the tree canopy. The understory maps produced represent effective input for the inversion of radiative transfer models (SAIL) and general useful information for forest ecological studies.