Pan Sharpening
Pansharpening is a process of merging high-resolution panchromatic and lower resolution multispectral imagery to create a single high-resolution color image. As we learned in previous sections, multispectral images have multiple bands and a higher degree of spectral resolution but often lack high spatial resolution. Panchromatic images on the other hand only one wide band of reflectance data but tend to have higher spatial resolution. The pansharpening process merges the multispectral and panchromatic images, which gives the best of both image types, high spectral resolution and high spatial resolution. Most high resolution imagery used in Google Maps and other mapping software is produced by pansharpening.
To apply pansharpening techniques images must either be georeferenced or have the same image dimensions. Ideally the multispectral and panchromatic images will be from the same sensor. Most newer multispectral sensors include a panchromatic band. Landsat 8’s Operational Land Imager (OLI) multispectral bands have a spatial resolution of 30m. Landsat 8 also collects a panchromatic band with 15 m spatial resolution. Pansharpening is ideal for improving the visual clarity of an image for interpretation or digitizing. The process does change the pixel values should it should be used cautiously for spectral analysis.
PanSharpening Techniques
There are several different methods used to pansharpen or merge images. Each method uses a different algorithm to merge the images. For example ENVI uses the following image sharpening techniques for:
- Gram-Schmidt Pan Sharpening
- Principal Components (PC)
- NNDiffuse Pan Sharpening
- Color Normalized (CN)
Pansharpening Example
