Resolution enhancement of hyperspectral imagery using maximum a posteriori estimation with a stochastic mixing model

A maximum a posteriori estimation method is developed and tested for enhancing the spatial resolution of hyperspectral imagery using higher resolution, coincident, panchromatic or multispectral imagery. The approach incorporates a stochastic mixing model of the underlying spectral scene content to develop a cost function that simultaneously optimizes the estimated hyperspectral scene relative to the observed hyperspectral and auxiliary imagery, as well as the local statistics of the spectral mixing model. The incorporation of the stochastic mixing model is found to be the key ingredient to reconstructing sub-pixel spectral information. It provides the necessary constraints for establishing a well-conditioned linear system of equations that can be solved for the high resolution image estimate.; The research presented includes a mathematical formulation of the estimation approach and stochastic mixing model, as well as enhancement results for a variety of both synthetic and actual imagery. Both direct and iterative solution methodologies are developed, the latter being necessary to effectively treat imagery with arbitrarily specified spectral and spatial response functions. The performance of the method is qualitatively and quantitatively compared to that of previously developed resolution enhancement approaches. It is found that this novel approach is generally able to reconstruct sub-pixel information in several principal components of the high resolution hyperspectral image estimate. In contrast, the enhancement for conventional methods such as principal component substitution and least-squares estimation is mostly limited to the first principal component.