A 3294*3012 IKONOS 4 bands image of Dubai coastline

home

go to work done in july 2009

work done in january 2009
 
 
 
 
 
 
Very gentle swell at shoreline, no sea-surface glint
Water appears quite homogeneous in Red and Nir bands, apart from some slightly turbid whirling plumes up to ~30 m away from shoreline seen in the Red band

 
 
 
Ikonos response curves
courtesy of S. Andrefouet

 
The following is a quest for operational wavelengths, using the seatruth dataset

It would seem that the green band behaves a bit like the Panchro band.

Blue band is well "peaked" at ~495 nm
NIR band is well "peaked" at ~780 nm
Pan band extends well below Red band

==>3-bands spectral bottom signature
may be obtained much deeper

 

 
  • The two round dots are an exposed sand bank , sampled by
    • -DN670/2283_5BrightBeach
    • -DN805/2174_5BrightBeach
  • Their radiances were introduced manually in the calibration data text file
    • AvShallow 2 + 670 2283 138.9 117.4 96.8 89.6
    • AvShallow 3 + 805 2174 151.4 136.4 119.4 122.3
  • Clearly,
    • ?LsM is grossly overestimated
    • the Soils Line should go through those round dots
    • Land pixels that were previously used to make up the SoilLine are fully loaded with adjacency effect
#nice -20 I start playing: cWL
#4SM.2.06 -PROCESS/31_JAN_2009/USF
-dB/PCI/dubaiIKYM/27_1_4_0/4_29/3294_3012/286947.479_2767543.207/1_1
-Origin/USF_dubai_4b.pix
-MIS/UAE/DUBAI/Ikonos/NA/USF/UTM_40_008/4_4/29_JUL_2001
-WLm444.5/506.4/631.8/757.3
-WLM515.9/595.5/697.7/852.7
-cWL00.50/00.95/00.50/00.50
-M@000001/0002/0003/0004
-Lsw078.7/46.5/22.2/16.1
-La0074.8/45.4/22.2/16.1
-LsM00200/0200/0200/0200
-Lm000001/00.5/01.8/0255
-LsT00001/0001/0001/0001
-LS000500/0500/0500/0500
-SCL00137/0195/0198/0180
-KK/1_2_0.760/7.0/400_500_600_700
-Z+/1.18/.1/0.0/0.0/1.0/PAN_1_3/SIG_3w5/N_1/1.0000
-B/tclNE_3.0/200_100/1/1.00 -LL10
-prepare/scale/delimb_1_40_51_10_Dr/fill_4_41/linearize
-deglintRBa/0.5/1,4_70/FN/1_22
-extractv/U16/4Max255/mBPL_1/mSOIL_21/mSHAL
-calibratev/BDSMpZ/4_3_2_1/bDH_12_23_34
-Model/mask_0 -Smooth_3d/Smart+
-RegressZZv/UTM-XYZ-sparse_GPS/0.001_0_0/1_0/ZM/2_3/0_201/0_30/*
@-DN670/2283_5BrightBeach

Seatruth on computed depth
using this calibration

Very poor color separation in the Blue/Green pair:

  • this points to coastal waters
  • modeling in the Blue/Green pair shall be noisy, to say the least
  • cWL[green]=0.95 ==> WL[green]=591 nm
  • Is this acceptable?








 
New LsM and La 

 
4SM.2.06 -PROCESS/31_JAN_2009/USF                 
-dB/PCI/dubaiIKYM/27_1_4_0/4_29/3294_3012/286947.479_2767543.207/1_1    
-Origin/USF_dubai_4b.pix                               
-MIS/UAE/DUBAI/Ikonos/NA/USF/UTM_40_008/4_4/29_JUL_2001                 
-WLm444.5/506.4/631.8/757.3                                             
-WLM515.9/595.5/697.7/852.7                                             
-cWL00.50/00.95/00.50/00.50                                             
-M@000001/0002/0003/0004                        
-Lsw078.7/46.5/22.2/16.1                        
-La0074.8/45.4/22.2/16.1                        
-La0068.0/46.0/22.4/16.1                        
-LsM00200/0200/0200/0200                        
-LsM00158/0140/0122/0100                        
-Lm000001/00.5/01.8/0255                        
-LsT00001/0001/0001/0001                        
-LS000500/0500/0500/0500                        
-SCL00137/0195/0198/0180                        
-KK/1_2_0.760/7.0/400_500_600_700                    
-Z+/1.18/.1/0.0/0.0/1.0/PAN_1_3/SIG_3w5/N_1/1.0000                         
-B/tclNE_3.0/200_100/1/1.00 -LL10                                       
-deglintRBa/0.5/1,4_70/FN/1_22                        
-extractv/U16/4Max255/mBPL_1/mSOIL_21/mSHAL                
-Calibratev/BDSMpZ/4_3_2_1/bDH_12_23_34     
             

CP arguments
 
The main improvement was obtained 
through rethinking of the Soils Line propblem
in the presense of a strong adjacency effect.




This describes/assumes 
that the waters become clearer 
where the bottom becomes deeper than ~3 m.
Is this likely at Dubai?

We all have memories of waters 
becoming clearer further away from shoreline.
 
nice -20 CP arguments
4SM.2.06 -PROCESS/31_JAN_2009/USF                
-dB/PCI/dubaiIKYM/27_1_4_0/4_29/3294_3012/286947.479_2767543.207/1_1   
-Origin/USF_dubai_4b.pix                              
-MIS/UAE/DUBAI/Ikonos/NA/USF/UTM_40_008/4_4/29_JUL_2001                
-WLm444.5/506.4/631.8/757.3                                            
-WLM515.9/595.5/697.7/852.7                                            
-cWL00.50/00.90/00.50/00.50                                            
-cWL00.70/00.90/00.35/00.24                                            
@CPf101.8/067.8/028.2/00000                       
-CPf097.8/063.8/026.2/00000                        
-CPf086.7/050.7/00000/00000                        
-CPf084.4/048.7/00000/00000       
                
-M@000001/0002/0003/0004                       
-Lsw078.7/46.5/22.2/16.1                       
-La0074.8/45.4/22.2/16.1                       
-La0068.0/46.0/22.4/16.1                       
-LsM00200/0200/0200/0200                       
-LsM00158/0140/0122/0100                       
-LsM/0176.0/0158.8/0141.9/0116.8_CoefLM=1.200               

-LsM/0176.0/0158.8/0135.9/0136.8                   
-Lm000001/00.5/01.8/0255                       
-LsT00001/0001/0001/0001                       
-LS000500/0500/0500/0500                       
-SCL00137/0195/0198/0180                       
-KK/1_2_0.760/4.0/400_500_600_700                   
-Z+/1.18/.1/0.0/0.0/1.0/PAN_1_3/sIG_3w5/N_1/1 .35000                        
-B/tclNE_3.0/200_100/1/1.00 -LL10                                      
-deglintRBa/0.5/1,4_70/FN/1_22                       
-extractv/U16/4Max255/mBPL_1/mSOIL_21/mSHAL               
-Calibratev/BdSNpZ/4_3_2_1/bDH_12_23_34                   
-Model/mask_0        -Smooth_3d/Smart+               
-RegressZZv/UTM-XYZ-sparse_GPS/0.001_0_0/1_0/ZM/2_3/0_201/0_30/*
    
Seatruth on computed depth
using new LsM, La and CP arguments

  
 
This is where I'm getting at
on Feb 2nd 2009.

Too much remains uncertain and sometimes puzzling
with the four Ikonos images I've had access to.

Too many questions unanswered:
  • Is the seatruth dataset tide-corrected?
  • Should I apply a tide height to my depths results when comparing with seatruth depths?
  • Which operational wavelengths for IKONOS imageries?
  • Why does the Blue band hardly correlate -if at all- with other bands in the atmospheric adjacency effect at Dubai  in july, while other bands are so tightly correlated?
  • By sampling the adjacency effect over the darkest patches, I obviously take the risk that such patches are not optically deep
  • There might even be negative bottom contrast in the Blue band to make things worse: this potentially obscures its correlation with other bands.
  • I seem to be twisting Jerlov's arms, and I certainly don't like that!!

 Seatruth on dark pixels only

 
Seatruth on bright pixels only

 
profile 1
 
profile 2
 
Profile 3

 
  
profiles 1, 2 and 3 from left to right



Créer un site
Créer un site