Evaluation of multiresolution image fusion algorithms

This paper is meant to mainly contribute to the comparative evaluation of image fusion algorithms used to merge the information content of spatially co-registered panchromatic and multispectral images. Multiple criteria and statistical indicators regarding the fidelity, spectral, spatial, and textural aspects of image quality are presented for objective and quantitative evaluation of the fused products for understanding the performance of image fusion algorithms. Experiments on merging QuickBird panchromatic and 3-band natural color multispectral images applying various fusion algorithms in spatial, spectral, and frequency domains indicates that the objective quality indicators presented in this paper perform significantly well for the evaluation of image fusion algorithms. Quantitative analysis shows that the FFS approach performs satisfactorily in preserving spectral fidelity and sharpening spatial and textural content.