Working with Landsat Data


Data Sources in order of ease of use

  1. University of Maryland's Global Land Cover Facility
  2. CaSiL Landsat Archive for CA
  3. USGS Global Viewer: free ortho-rectifed imagery

Sample GRASS commands

#display a true color composite image from Landsat TM data:
d.rgb r=landsat_band_3 g=landsat_band_2 b=landsat_band_1
#fuse bands into a single image:
r.composite r=landsat_band_3 g=landsat_band_2 b=landsat_band_1 out=landsat_bands_321
#blend landsat composite image with DEM data
#note that blending will require some tweaking: good results obtained with:
r.blend second=shaded_relief first=landsat_321 out=landsat_shade perc=6
#re-composite the blended results
r.composite landsat_shade.r g=landsat_shade.g b=landsat_shade.b out=landsat_shade.rgb
#create histogram equalized inputs (Fig 5)
r.colors map=landsat_band_1 color=grey.eq
r.colors map=landsat_band_2 color=grey.eq
r.colors map=landsat_band_3 color=grey.eq
r.composite r=landsat_band_3 g=landsat_band_2 b=landsat_band_1 out=landsat_eq_321

Simple comparison between the default output from r.composite on landsat bands 3,2,1 and the enhanced output from i.landsat.rgb.

Composition of Landsat TM bands 3, 2, 1 (red, green , blue)
Fig.1 Bands 3,2,1
Composite of Landsdat TM bands 4, 3, 2
Fig.2 Bands 4,3,2
Composite of Landsdat TM bands 5, 3, 2
Fig.3 Bands 5,3,2
2x exaggeration on relief, 6x weighting of landsat bands 3,2,1
Fig.4 Blend with hillshade
Histogram equalized 321
Fig.5 Histogram equalized 321
Default output from r.composite of bands 3,2,1
r.composite default
Composite from AZ created with i.landsat.rgb, c/o Markus Neteler
i.landsat.rgb enhanced