Sunday, January 31, 2016

Trying out NNDiffuse Pan Sharpening in ENVI 5.2

Its been a while since last time I am exploring pan sharpening. Recently I am stumbled on ENVI which had  just released version 5.3 (version 5.4 will be released around summer 2016 as far I know) and it is brings some new tools in their arsenal. One of them is NNDiffuse Pan Sharpening Algorithm. ENVI has known for very long time only have 4 common image fusion/pan sharpening algorithm (PCA, Brovey, CN, and IHS/HSV) thats not quite good to maintain spectral quality of the fusion/pan sharpening result. At later version, they are adding Gram Schmidt Spectral Sharpening which is one of the most advanced pan sharpening algorithm to date and making them equal with ERDAS (Ehlers/wavelet fusion) and PCI (UNB Fusion) as the most advanced Pan Sharpening Software. Now they are adding NNDiffuse which is the latest algorithm developed by Weihua Sun et al,


NNDiffuse utilizing regression approach combined with spatial frequency extraction to generate pan sharpened dataset which not only maintain the spectral quality of multispectral bands, but also well preserve the spatial details of panchromatic band. And more of it, this algorithm also works for SWIR bands which its spectral range are beyond the spectral range of panchromatic band (similar with Gram Schmidt or other frequency based pan sharpening algorithm which also works for SWIR Bands). 


So I have been recorded my NNDiffuse testing using Landsat-8 Data into a video tutorial which you can check it out below.