Bathymetry and water column correction
WorldView 2 image Princess Cays, Eleuthera island, Bahamas
  November 11th 2011, 2 m ground resolution
image from Digital Globe
"(c) DigitalGlobe, Inc. All Rights Reserved"
"Includes copyrighted material of DigitalGlobe, Inc., All Rights Reserved"
work done in 2012, 2013 and 2014
The "PAN solution"
uses the PANCHRO bans, 4*4 sum-aggregated and co-registered
bands 1, 2, 3, Panchro, 4, 5, 6, 7 and 8       become      bands 1, 2, 3, 4, 5, 6, 7, 8 and 9
home
 
 
 
Profiles

Porfile_Black location		 Location of Profiles
 
  • This work uses bands Purple, Blue, Green, PAN, Yellow and Red
  • Use of the PAN band yields excellent results:
    • it would seem that its optical calibration is well under control

 
In the plots below:
  • Red    is computed depth dated Sept 28th 2012
  • Black is computed depth dated Oct   20th 2013
Profile_Black
using all bands
  • in section B, from 1.7 km to 3.0 km, depths are computed using Purple, Blue, Green and PAN bands
  • in section A, from 0.7 km to 2.1 km, depths are computed using Purple, Blue, Green and PAN bands
  • Depth result as of Sept_25th is plotted in red for comparison

 
In the plots below:
  • Red    is computed depth dated Sept 28th 2012
  • Black is computed depth dated Oct   20th 2013
Profile_Black  All_PAN
not using Red and Yellow bands might even be better
  • all depths are computed using Purple, Blue, Green and PAN bands
  • Depth result as of Sept_25th is plotted in red for comparison

Profile_Red
using all bands
  • Depth result as of Sept_25th is plotted in red for comparison

 
Profile_Red  AllPAN
not using Red and Yellow bands might even be better
  • Depth result as of Sept_25th is plotted in red for comparison

 



 

Modeling results
the "PAN solution"
uses purple, blue, and green bands against the PAN band over the whole scene
this yields in a seamless result (no gaps)
 

Z in dm
 
Z in cm
 

TCC
TCC water column corrected
looks rather dull

image B
this is a colorised view
of the water column corrected blue band
TCC watercolumn corrected, normalized


 

Bottom typing
there is good substance for bottom typing

Normalized Difference Index

NDR=(LBpurple-LBgreen)/(LBpurple+LBgreen)
LB is the BOA spectral
water column corrected bottom reflectance
The NDR image is computed "on the fly"
Green colors for NDR<0
Purple colors for NDR>0
TCC watercolumn corrected, normalized

This is sort of a weird enhancement
of the water column corrected TCC
It is computed "on the fly"





 

Conclusions
 
Using the Panchromatic band
  • Looks like problem solved
  • Chances are that the accuracy of retrieved depth has improved significantly
  • Use of PAN band would appear to be under reasonable control
    • including at very shallow depth
    • so we can dispense with using Yellow and Red bands
  • Still, the optical calibration of the PAN band needs to be refined, in order to reduce the high level of noise in depth results: needs some more time and efforts, as the problem is not quite simple to solve
  • All these novelties are hard coded into 4SM code
    • based on work on WV2 at Perth and Princess Cays, OLI at SharkBay, Marawah, and then WV2 Princess Cays again
  • Now, I want to use the DTMs for fine tuning of all these novelties
    • and would certainly want to test them on new quality WV2 images
Using the PAN band might be of particular interest in coastal waters
where Purple, Blue and Green bands have virtually the same attenuation coefficient,
which means: there is no color separation deeper than a very few meters
THIS IS WORTH TRYING ASAP

I'd want to see if I can use a PAN_sharpened image:
looks like YES
that would be very interesting for Landsat 8 OLIP images
 

The command line script, as of October 2013
nice -20  4SM  -Process  -Origin/DigitalGlobe                       
-DB/princesscayswvmp/50_3S_9S_0s/9_54/6001_5701/373.655_2732.957/1_1           
-Mis/Bahamas/PrincessCays/WV02/MultiPan/PIX/UTM_18_008/0.002_0.002/11_NOV_2011       
-LS/005000/05000/05000/05000/05000/05000/05000/05000/05000               
-WL/0427.0/477.5/546.0/570.0/607.5/679.0/724.0/831.0/908.0       
-LsM/200.2/192.0/187.5/215.4/202.5/210.4/210.3/180.7/175.6       
-Lsw114.01/74.15/38.23/30.46/25.14/17.89/13.39/07.13/06.08_glint   
-dLsw/0.20/00.20/00.00/000.5/0-.10/00.25/000000/0000/00000       
-Lw/0013.0/008.3/000.5/0011.5/000.5/000.0/000.0/000.0/000.0       
-SCL/00095/00107/00167/02500/00171/00116/00166/00164/00149       
-Lm/0000.5/000.5/006.3/000.2/002.0/002.0/001.0/001.0/255.0       
-M/@000001/@0002/00003/00004/@0005/@0006/@0007/@0008/00009       
-KK2_3_0.5500/Knir8.900/mask_3                       
-Z/MSL0.00d/N_2/nDR_53_zDTM/mask_3                   
-B/TclNe/cNorm5.00/cLM1                         
-deglint/vRbaD/Q/F1/L9/GlintM25.0                   
-extract/v/rawBDH/fullBDH/NIRband8/NIRmax5/mapBPL/mSOIL21/mBPL2       
-calibrate/V/BdSMpZg/plot_BPL/BC_6_5_4_2/BDh_12_23_34            
-Model/v/runWV2/mask_3  -TPs/2/4/Text/cPAN/A6.3615_B-0.0192_C0.0052_D2 
-Smooth/5/D/Smart+
 

Download the results in georeferenced TIF files (134 MB)
princesscaywv_results_21oct2013.zip
from GoogleDrive
it contains:
  • princesscayswvmp_Zcm.tif.......................................................................16 bits signed
    • Retrieved depth in centimeters
  • princesscayswvmp_NDR.tif .....................................................................16 bits signed
  • princesscayswvmp_WCC.tif: water column corrected bands....................8 bits
    • LB...............................channels 1 to 6...........for purple, blue, green, panchro, yellow and red
    • Normalized TCC...........channels 7, 8 and 9.....for blue, green and red
 



Créer un site
Créer un site