Comparison of Fusion Algorithms for ALOS Panchromatic and Multispectral Images

Image fusion techniques of remote sensing data can integrate different spatial and spectral information. ALOS is a new data source with two optical sensors onboard, PRISM and AVNIR-2. PRISM has high spatial resolution, while AVNIR-2 is of multi-spectral resolution. Study on ALOS data fusion methods has potentials for promoting the applications of ALOS data. This paper aimed to explore a proper method for data fusion purpose. Five fusion methods, Brovey, PCA, SFIM, HPF and Gram-Schmidt, were tested. Visual and statistical comparisons were conducted to evaluate the effect of different fusion algorithms in terms of spatial resolution enhancement and spectral information maintenance. We found overall results of Gram-Schmidt in improving spatial resolution details and preserving spectral information were superior to the others. PCA was good at enhancing textural information, but caused much color distortion. SFIM best preserved the spectral abundance, but was less optimal at improving spatial information.