The approach we propose is based on the development and combination of physical and statistical models for both observations and prior knowledge and the subsequent use of these models for optimal or near-optimal processing.
Feature
extraction: dimensionality reduction in hyperspectral
imagery, texture analysis in multi-dimensional imagery;
Physical modeling and inverse problems: microwave remotes sensing of the ocean and the atmosphere, fast codes for reflectance to radiance conversion, and clutter modeling.
Data fusion: atmospheric and ocean parameter estimation, resolution enhancement, detection and discrimination, and multimodal data fusion.
Data management: lossless compression, compression of multispectral and hyperspectral imagery.
