Efficient hierarchical fusion using adaptive grouping techniques for visualization of hyperspectral images

Visualization of hyperspectral images that combines the data from multiple sensors is a major challenge due to huge data set. An efficient image fusion could be a primary key step for this task. To make the approach computationally efficient and to accommodate a large number of image bands, we propose an efficient hierarchical fusion based on adaptive grouping techniques. The consecutive image bands in the hyperspectral data cube exhibit a high degree of feature similarity among them due to the contiguous and narrow nature of the hyperspectral sensors. Exploiting this redundancy in the data, we generate clusters of similar images and fuse them at every level of hierarchy using bilateral filter. For grouping similar images, we propose six very generic yet proficient similarity criteria based on either information theoretic metrics or image quality metrics. We also investigate each similarity criterion based on statistical evaluation of quality of final fused image.