NAME

i.colors.enhance - Performs auto-balancing of colors for RGB images.

KEYWORDS

imagery, RGB, satellite, colors

SYNOPSIS

i.colors.enhance
i.colors.enhance --help
i.colors.enhance [-fprs] red=name green=name blue=name [strength=float] [--help] [--verbose] [--quiet] [--ui]

Flags:

-f: Extend colors to full range of data on each channel
-p: Preserve relative colors, adjust brightness only
-r: Reset to standard color range
-s: Process bands serially (default: run in parallel)
--help: Print usage summary
--verbose: Verbose module output
--quiet: Quiet module output
--ui: Force launching GUI dialog

Parameters:

red=name [required]: Name of red channel
green=name [required]: Name of green channel
blue=name [required]: Name of blue channel
strength=float: Cropping intensity (upper brightness level)
Options: 0-100
Default: 98

DESCRIPTION

i.colors.enhance auto-balances and enhances the color channels of a RGB image (e.g. from Landsat) to provide a more natural color mixture. Only the color table of each image band is modified, the base data remains untouched.

The module works by calculating a histogram for each color channel and removing an adjustable amount of outliers from either end before recalibrating the color scale with r.colors.

It will work with any 8-bit RGB imagery set and the script is easily modified to work with other datasets of greater band-depth.

NOTES

Depending on the image, it may or may not be advantageous to use the -p flag to preserve the relative color scaling. You will have to experiment with the different options to find a setting that works best for your particular imagery.

The strength option should generally be set in the 90-99 range. The lower the number, the more saturated the image becomes. It represents the percentage cut-off for the top end of the color histogram curve. The lower end is fixed at 2% of the area under the curve.

For quicker execution of this module on large images you can achieve largely similar results by switching to a coarser resolution before the running of the module (using g.region) and then back to the original resolution afterwards.

EXAMPLE

North Carolina sample dataset example with Landsat data:


g.region raster=lsat7_2002_10 -p
d.rgb blue=lsat7_2002_10 green=lsat7_2002_20 red=lsat7_2002_30
d.barscale at=55.3125,93.125 bcolor=white text_pos=over
i.colors.enhance blue=lsat7_2002_10 green=lsat7_2002_20 red=lsat7_2002_30 strength=95
d.rgb blue=lsat7_2002_10 green=lsat7_2002_20 red=lsat7_2002_30
d.barscale at=55.3125,93.125 bcolor=white text_pos=over

Uncorrected RGB composite of Landsat ETM channels B/G/R

Color corrected RGB composite of Landsat ETM channels B/G/R

TODO

The strength option requires further refinement.

AUTHORS

Markus Neteler, Trento, Italy
M. Hamish Bowman, Dept. Marine Science, Otago University, New Zealand

SOURCE CODE

Available at: i.colors.enhance source code (history)