Bathymetry and water column correction
LANDSAT 8 OLIP 
at SanLorenzoChannel, Baja california
Image courtesy of the U.S. Geological Survey
7671*7841 30 m pixel size, UTM zone 12
 
Using the Panchromatic band for water column correction
to derive water depth and spectral bottom signature:

Landsat 8 OLIP bandset used for this work
Purple=1, Blue=2, Green=3, PAN=4, Red=5, NIR=6 and SWIR1=7

scene LC80340432014007LGN00, January 07th 2014
Work done october 2016
home
 
 
GREEN solution or PAN solution?
15m resampling or not?
15 m PAN-sharpening or not?
Smoothing or not?

One more trick: apply ZDTM






GREEN solution or PAN solution?
GREEN solution
is used by most practioners
after atmospheric correction
and calibration using depth soundings
PAN solution
is a breakthrough by 4SM
requires neither atmospheric correction
 nor calibration using depth soundings
 
  • cover the whole depth range, a little deeper than the PAN solution
  • only returns Coastal, Blue, Green water column corrected bands
  • is very sensitive to proper estimation of water volume reflectances: LwGREEN is far from negligeable
  • is highly dependent on the SOIL LINE assumption
  • bottom typing: limited to bottom brightness
  • commonly causes frustration
    • where water is not very clear: not for Coastal waters, requires a low value for ratio KBBLUE/KGREEN
    • over green/dark substrates and at very shallow depth, as most practioners do not bother with water volume reflectances
  • covers the whole depth range, a little less than the GREEN solution
  • returns Coastal, Blue, Green and PAN water column corrected bands
  • is much less sensitive to proper estimation of water volume reflectances: LwPAN is allways very low
  • bottom typing: yields a more contrasted bottom signature where the GREEN band is much less limited by the SOIL LINE assumption
    • this is good for SAM bottom typing
  • delivers much better results over green/dark substrates and at very shallow depth
  • is more noisy at depth
I prefer PAN solution
GREEN solution impractical in this water






15m resampling or not?
30 m GSD
15 m GSD
requires resampling the PAN band
from 15m GSD to 30m GSD		 requires resampling the MULTI bands
from 30m GSD to 15m GSD
 
  I prefer 15 m GSD resampled





15 m PAN-sharpening or not?
4SM or R?
15 m GSD resampled
  • for each pixel, the PAN radiance is operated against a 15 m resampled carbon copy of the original MULTI radiance spectrum
  • this results in a somewhat "blocky"  fabric noise in the results
    • which may warrant a light smoothing scheme
15 m GSD pan-sharpened
  • for each pixel, the PAN radiance is operated against a 15 m pan-sharpened avatar of the original MULTI radiance spectrum
  • this is the cause of a high level of noise in the results which may be detrimental
I prefer 15 m GSD resampled BUT:
looks like your own two pan-sharpened images
gave me quite some satisfaction:
I need to go back, see why that!
Maybe the R routine is much better than mine



Smoothing or not?
by a 5 pixels circular kernel
Pink bottoms
No smoothing applied
  • this is required for ground truthing purpose
    • regressions, RMSE and the like
Smoothing applied
  • this is required for practical purposes:
    • bathymetric maps
    • bottom typing maps
see GSD 30 m Not smoothed    Profile Black
 Profile Black   see GSD 30 m Smoothed
 
 
NOTE:
Pinkish hues in SAM signal 
are not an artifact
They actually cover a wide range of bottom brightness, which includes very bright (saturated) bottoms
What causes it?
I am still looking for your rhodoliths!

 see legend for SAM
Some SAM bottom signatures
SAM 10 is reference bright bottom
SAM 12, 14 etc are a suite of greenish signatures
SAM 31, 33, 35 etc are a suite of pinkish bottoms




 



Créer un site
Créer un site