Pan Sharpening is an analysis method by which lower resolution color imagery can be sharpened (increased in resolution) using a higher resolution panchromatic image. This tool can be found in the Raster Analysis menu.
Input Color image and Panchromatic image, produce a Pan-Sharpened Image (IHS) .
Pan Sharpening is typically done with remotely sensed imagery, where a high resolution panchromatic image was collected simultaneous to lower resolution color imagery. For example Landsat8 and Spot5 both collect high resolution Panchromatic bands. The required inputs to perform pan sharpening are a single band high resolution image and a lower resolution 3 or 4 band image (typically RGB or RGBI).
Algorithm
Specify the algorithm to be used for Pan Sharpening. Each method takes every pixel from the panchromatic image (or as specified in the sample spacing) and determines a value for each band using the color image. The band values are normalized based on the image histogram (min and max values) and any contrast adjustment settings that have been applied. The algorithm used will be stored in the metadata of the pan-sharpened image.
-
Brovey - In the Brovey Method, each band in the color image is multiplied by the panchromatic band and divided by the sum of the color band values. This divides the value in the panchromatic band across the color bands using a ratio calculated from the proportions in the original color image. This method increases contrast in the light or dark color tones.
- ESRI Transform - This method adjusts the original color band values. The adjustment is calculated by determining the difference between the panchromatic image value and a weighted average of the color bands.
-
IHS (Simple Intensity Replacement) - This is the initial default algorithm. The color image bands are converted to intensity, hue and saturation values, then after histogram matching, the intensity values of the color band are replaced with the panchromatic intensity values. -
Simple Mean - The simple method calculates the mean between the color band value and the panchromatic value in order to produce each output color band.
The resampling method for determining the values used in the algorithm at each output pixel location come from the layer options. For imagery the default setting is Nearest Neighbor.
Color Layer
Specify the color image to pan-sharpen. This can be a 3 or 4 band multi-band image, or an RGB image.
Pan Layer
Specify the panchromatic layer. This must be a single band image. Typically pan sharpening is performed with a panchromatic image collected simultaneous to a color image.
Panchromatic imagery is collected by sensing all wavelengths of visible light, as opposed to color imagery where bands are collected separately based on the wavelength.
Band Weights
Band weights are factored into the chosen algorithm. Each method uses a sum or average calculation of the RGB (I) color band values when calculating the output bands. Different weighting can be applied by choosing to use custom band weights. The band weight values should add up to 1, however compensation will be done if they do not.
Band weighting can be applied to the specified pan sharpening algorithm to correct color, or enhance certain parts of the image.
Comments
0 comments
Please sign in to leave a comment.