Oft-normalize.bash

From Open Foris Wiki

Jump to: navigation, search

This script normalizes image grey values on each band and training data file


Usage: oft-normalize.bash <-i image> [-t training data] [-f 1/2 ] [-m mask]

Options:

  • -i image = give the Landsat image with 6 or 7 bands to be normalized
  • -t training data = give a text file containing ground truth and image bands (in last columns)
  • -f 1/2 = normalization will be based on the distribution present in the image (1) or the training data file (2)
  • -m mask = give a mask file showing areas to be processed with 1 and others with 0


  • Image grey values in both files are converted to mean 0 and std 1 based on the selected source of distribution (image or training data file)
  • Procedure for converting each grey value on each band in the image and/or training data file is (value - average)/std
  • It is possible to
  1. Normalize just the image based on it's grey value distribution on each band
  2. Normalize also the training data text file using the same distribution or
  3. Normalize both files using the grey value distribution obtained from the training data file


Exercise:

  • Get Example data set
  • For this exercise following tools are used: oft-normalize.bash
  • Open your working directory using
cd /home/...
  • Let's run a simple exercise using landsat_t1.tif as the only input:
oft-normalize.bash -i landsat_t1.tif

Output: landsat_t1_norm.tif & stat_landsat_t1.txt

  • Now we run it including the training data option values_for_nn:
oft-normalize.bash -i landsat_t1.tif -f values_for_nn



Back to Open Foris Toolkit Main Page

Back to Tools & Exercises



Personal tools