Arcachon tutorial: Calibration 2
back to Arcachon tutorial      back to Model 1        next to Model 2
 
 
 
Here is our last command line
#Line_3 Model 1+++++++++++Model 1+++++++++++Model 1+++++++++++
nice -20  ./4SM.4.08 -Process/14_Jul_2011/06_JUL_2011  -OriginSPT @@
-DB/arcachon/27_1_0_0/3_1/500_630/0.010_22.610/1_1 @@
-Mis/Arcachon/France/Spot/XS/NA/UTM_1_008/0.020_0.020/1_AUG_1986 @@
-LS/0255.0/255.0/255.0 @@
-M/@000001/00002/00003 @@
-LsM/078.8/071.2/065.2_cLM=0.450 @@
-Lsw/030.4/014.2/008.0 @@
-Lw/0001.5/000.0/000.0 @@
-Lm/0001.0/001.0/255.0 @@
-KK1_2_0.3900/Knir4.500/mask_1 @@
-Z/MSL0.00/cZ1.000 @@
-B/tclNe/LBref200_100/Bmin0/cLM1.00 @@
-extract/v/mBPL2 @@
-calibrate/v/BDSNpzg/BDh_12_13 @@
-Model/mask_2 @@
-Smooth/5/D/Smart+ @@
@Recode/201_InChannel_9_into_233_OutChannel_4

Now, we want to improve/consolidate the optical calibration
to get rid of the breaking wave pixels in the BPL
First get a "clean" calibration data textfile

Please copy Line_3 to Line_4

 
#Line_4 Calibration 2+++++++++++Calibration 2+++++++++++Calibration 2+++++++++++
nice -20  ./4SM.4.08 -Process/14_Jul_2011/06_JUL_2011  -OriginSPT @@
-DB/arcachon/27_1_0_0/3_1/500_630/0.010_22.610/1_1 @@
-Mis/Arcachon/France/Spot/XS/NA/UTM_1_008/0.020_0.020/1_AUG_1986 @@
-LS/0255.0/255.0/255.0 @@
-M/@000001/00002/00003 @@
-LsM/078.8/071.2/065.2_cLM=0.450 @@
-Lsw/030.4/014.2/008.0 @@
-Lw/0001.5/000.0/000.0 @@
-Lm/0001.0/001.0/255.0 @@
-KK1_2_0.3900/Knir4.500/mask_1 @@
-Z/MSL0.00/cZ1.000 @@
-B/tclNe/LBref200_100/Bmin0/cLM1.00 @@
-Extract/v/mBPL2 @@
-Calibrate/v/BDSNpzg/BDh_12_13 @@
-Model/mask_2 @@
-Smooth/5/D/Smart+ @@
@Recode/201_InChannel_9_into_233_OutChannel_4
 
please note that the -Extract... and/or the -Calibrate... arguments take precedence over other actions

Now run the command line script
  • Note that the Darkest Pixels Line is plotted   using cross symbols
    • Remove it if you like : -Calibrate/v/BdSNpZg/BDh_12_23       
  • Note that the Z2 vs Z3 insert is present
    • Remove it if you like : -Calibrate/v/BdSNpzg/BDh_12_23       
 

Before
 

After
 

Before
most BPL pixels are wave breaker pixels

 

After
all BPL pixels are genuine shallow pixels

 

 
 
  • Deeper than ~0.1 m, 
    • the Red_vs_NIR BPL model using KNIR=4.5 m-1 fits quite neatly the selected BPL pixels
  • The Green_vs_NIR BPL pixels are quite messy though
Where a blue band is available, it observed that
  • Red_vs_NIR is quite good compared to Green_vs_Nir and Blue_vs_Nir
    • allows a good estimation of the ratio Kred/KNIR 
  • Green_vs_Red is much better than Blue_vs_Red
    • allows a good estimation of the ratio Kgreen/Kred Kgreen/Kred
  • Blue_vs_Green is quite good
    • allows fairly good estimation of the ratio Kblue/Kgreen Kgreen/Kred
By application of  Kgreen/Kred =0.39
  • Water type is clearly Coastal 1+0.8 of Jerlov : almost Coastal 2 : quite neat!
    • K550=0.359 m-1
    • K475=0.522 m-1
    • the "effective wavelength" is ~550 nm : this makes for "green" waters
    • a blue band would not help!!!!!!!
  • KNIR=4.5 m-1 is gives a nice confirmation of this calibration of the optical properties of the waters at Arcachon
  • still quite a few BPL pixels are questionable in the Green vs Red pair of bands
 

 
Things you can do
See if Smoothing makes any difference
  • -smooth/5/D/Smart+ 
Notice that the brightness of most sandy pixels is only 0.70 that of the BPL pixels
  • -B/tclNe/LBref200_100/Bmin0/cLM0.70
Notice that Kgreen/Kred =0.42 achieves a better fit of the BPL model
  • Coastal 2 water type            -KK1_2_0.42/Knir4.500/mask_1
  • Coastal 2 water type            -KK1_2_0.39/Knir4.500/mask_1
Notice that Lwgreen=2.5 would be more appropriate
  •   Lwgreen=1.5 was assigned by default in the AutoCal process
    • this places the average vegetation (blue dot) right under the Soil Line in the X1 vs X2 diagram 
  •  Lwgreen=2.5   achieves a physically more consistent scenario:
    • where the blue dot plots on the Green side of the Soil Line,
    • as may be expected,
    • considering that the reflectance of the vegetation is stronger in the Green band than in the Red band. 

Green vs Red, linearized
  • Three isobath lines in this plot are: 0, 2 and 4 m
    • maximum retrieved depth is ~  6 m for the Red           solution over bright bottoms
    • maximum retrieved depth is ~15 m for the One-band  solution over bright bottoms

#Line_4 Calibration 2+++++++++++Calibration 2+++++++++++Calibration 2+++++++++++
nice -20  ./4SM.4.08 -Process/14_Jul_2011/06_JUL_2011  -OriginSPT @@
-DB/arcachon/27_1_0_0/3_1/500_630/0.010_22.610/1_1 @@
-Mis/Arcachon/France/Spot/XS/NA/UTM_1_008/0.020_0.020/1_AUG_1986 @@
-LS/0255.0/255.0/255.0 @@
-M/@000001/00002/00003 @@
-LsM/078.8/071.2/065.2_cLM=0.450 @@
-Lsw/030.4/014.2/008.0 @@
-Lw/0001.5/000.0/000.0 @@
-Lw/0002.5/000.0/000.0 @@
-Lm/0001.0/001.0/001.0 @@
-KK1_2_0.420/Knir4.500/mask_1 @@
-Z/MSL0.00/cZ1.000 @@
-B/tclNe/LBref200_100/Bmin0/cLM1.00 0.70 @@
-Extract/v/mBPL2 @@
@Extract/v/RawBDH/NIRband3/NIRmax5.0/FullBDH/mSOIL21/mBPL2 @@
-Calibrate/v/BdSNpzg/BDh_12_13 @@
-Model/mask_2 @@
-smooth/5/D/Smart+ @@
@Recode/201_InChannel_9_into_233_OutChannel_4

 



 



Créer un site
Créer un site