This code will take a RapidEye Image and apply a 2 standard deviation color balance. It will then rescale the image to 8bits and save a new TIF.
Purpose: Do a crude color balance on several images so we can create a 'color balanced' mosaic.
I explained this in the post:
https://gis.stackexchange.com/questions/95435/color-balance-and-publish-large-amount-of-high-resolution-imagery-in-geoserver/100097?noredirect=1#comment496785_100097
Script will run on one image at a time. So my recommendation is to use it with xargs, in order to do a parallel processing.
Example run:
-
Generate a list of files to process
grep -f lista_pr.txt todos_arquivos_tif_2014.txt > todos_arquivos_2014_pr.txt -
Run with
xargs
cat todos_arquivos_2014_pr.txt | xargs -n 1 -P 5 -I % /c/projetos/codigos/rapideye/rapideye_hist2sd.py -b 1 2 3 4 5 -m -md /z/rapideye2014_udm -sk 5 % .
for large file lists, it might be better to split it in smaller chunks. Take a look at split command
-
Get a list of the output images and create a mosaic
ls 8bit*.tif > lista_mt2014_normalizado
gdalbuildvirt.exe -b 5 -b 3 -b 2 -input_file_list <lista dos arquivos> <rs2014_b532.vrt> -
Convert to TMS or whatever format you wish
gdal2tiles.py -z '5-15' rs2014_2sd_udm/rs2014_b532.vrt rs2014_b532_tms
Some before / after images:
Before...
Brazil Southeast, no color balance
Brazil Southeast, no color balance - zoom
After...
Brazil Sul mosaic, crude color balance
Brazil Sul mosaic, crude color balance - zoom