WorldView 2 image Princess Cays, Eleuthera island, Bahamas
  November 11th 2011, 2 m ground resolution
  "(c) DigitalGlobe, Inc. All Rights Reserved"
"Includes copyrighted material of DigitalGlobe, Inc., All Rights Reserved"

home

This dataset from Digital Globe under Data Exchange Agreement
courtesy of Gregory MIECZNIK: watch his bathymetry webinar
 
 
 

Final MULTI calibration september 2014
Bandset has changed: I finally managed to read/add the Panchromatic band as channel 4
bands 1, 2, 3, Panchro, 4, 5, 6, 7 and 8        
become      bands 1, 2, 3, 4, 5, 6, 7, 8 and 9

 
Seatruth on work done Sept 28th 2012 is available: I have worked knowing that :
  • 0-8 m depth range : minor adjustments are required
  • 8-30 m depth range : major adjustments are required
  • preliminary calibration suggested it anyway
New calibration tools now allow for fairly secure choice 
of operational wavelength for the Green  band : 

this is a very important step forward for 4SM,
and it clarifies confusing findings obtained when using Lansdat imageries
Final MULTI calibration september 2014
after shifting the effective wavelength for the Green band
PAN band still not in use

 
THE PURPOSE IS 
to achieve the best possible fit in the calibration diagrams
of the observed scatter plot of BPL pixels
with the various ratios Ki/Kj which statisfy the rule
  Ki/Kj=(Ki/Kk)/(Kj/Kk)
AND THEN
to apply Jerlov's data to derive spectral K in the visible range 
SO THAT : NO NEED FOR FIELD DATA!
 

Final calibration
for bands 2, 3, 5 and 6
-WL475.5/477.5/557.9/570.0/607.5/659.0/724.0/831.0/908.0
  • Wavelength for the Green band 3 has been increased from 546.0 nm to 557.9 nm
    • Kblue was 0.074 m-1 Kgreen was0.142 m-1
    • now Kblue is 0.091 m-1 Kgreen is 0.174 m-1
    • this is a major change: as a result, all retrieved depths shall be decreased by a factor of 0.81
  • Ratio K2/K3=0.52 remains unchanged
  • Water type has shifted
    • from OIB+0.104 water type of Jerlov
    • to     OIB+0.394 water type of Jerlov
  • See that, as a result of this small change, a good fit is now observed for the Yellow and Red bands in this final calibration:
    • all computed depths should now be reduced by a factor of 0.142/0.174=0.81:
      • that's ~1.5 m shallower at 8 m
      • and ~5 m shallower at 25 m
    • this would appear to be compatible with Gregory's seatruth plot of 4SM retrieved depths
  • FIRST ORDER: therefore, increasing the effective wavelength of the green wideband would appear to be the way to go in 4SM modeling
    • it has a first order effect on the estimated Jerlov water type
    • then YES: NoNeed For Field Data
  • This needs confirmation though, as there might be a dependence on the structure of the light field at the time of imaging

Final calibration
for bands 1, 3, 5 and 6
-WL475.5/477.5/557.9/570.0/607.5/659.0/724.0/831.0/908.0

 
  • Wavelength for the Coastal band 1 had to be be increased manually from 427.0 nm to 475.0 nm
  • This is a major change with a minor effect
  • See that, as a result of this change, a much improved fit is now observed for the Coastal band 1
    • but this change is more difficult to justify
 
  • Wavelengths in the Yellow and Red bands remain unchanged
    • only the wavelength for the Green band has been increased until a good fit is observed for the Yellow and Red bands
  • Kyellow has increased from 0.535 to 0.545 m-1
  • Kred       has increased from 0.799 to 0.814 m-1
  • These are minute changes,
    • which should prove hardly noticeable in the depth results,
    • but show off distinctly in a color composite of water column corrected radiances
  • Therefore, shifting the effective wavelength of the Green band is the only thing that needs to be done:
    • it has a major effect on depths computed in the 8-30 m depth range,
    • and a minute effect on depths computed in the 0-8 m depth range




Final MULTI Profiles september 2014
after shifting the effective wavelength for the Green band
PAN band still not in use

In the following profiles, the september 28th 2012 has been used as "seatruth" and drawn in red
See how retrieved depths have decreased in the 8-30 m depth range

Profile Red on the western side
where seagrass abund
Note that waypoints have been placed to highlight the vertical banding which causes a bad depth artifact

Profile Purple on the western side
where seagrass abund
This is the location of UKHO's MBES data,
right in front of Princess Cays resort

Profile Yellow on the eastern side
where corals abund
  • the situation remains confuse
  • this shall clear out by using the PAN solution

Profile Green on the eastern side
where corals abund
  • the situation remains confuse
  • this shall clear out by using the PAN solution

 
 
 



Créer un site
Créer un site