4SM SPOT tmnov-tutorial-calibration5

Bi-dimensional histogram for marine pixels only

Full BDH was turned ON
- -Extract/v/FullBDH/mBPL2 extracts the BiDimensionalHistograms of all pixels = the default
Now, turn it OFF
- -Extract/v/fullBDH/mBPL2 extracts the BiDimensionalHistograms of marine pixels only

Calibration 5 : FULL BDH
BiDimHistogram shows land and marine pixels

Calibration 5 : shallow BDH
BiDimHistogram shows only marine pixels

Calibration under a local mask

What of the Soil Line?

Look at the outer reef under mask_4

First================================================

-Recode/2_InChannel_5_into_4_OutChannel_5/mask_4
In channel_5 : recode to 4 all pixels under mask_4.shp that are masked at 2

Then================================================

-Extract/v/fullBDH/mBPL4 f : marine pixels only mBPL4 : under mask_4
-Calibrate/v/BDSNpZg/BDh_12_23_34 D : shall also show the DPL pixels

Calibration 5 : whole shallow BDH
BiDimHistogram shows only marine pixels
for the whole image

A great number of BiDimHistogram pixels
- are located on the right side of the Soils Line:
They can't be assigned a computed depth,
- unless we re-consider the concept of SoilLine.
-Z/MSL0.00/RLBgb1.250/cZ1.000
RLBgb1.25 is there to include that aspect in 4SM
- this states that, when modeling by the 2-bands case, the "SoilLine" is now a straight line that runs through LsM with a slope=1.25 in normalized data
- In other words,
  - the darker the bottom reflectance, the greener the computed spectral signature
  - the higher RLBgb, the higher the computed depth
This of course is not fair to any pixel whose actual bottom signature would not be greenish
But, as we shall see later, it suits the vast majority of shallow pixels at Caicos Bank, actually everywhere round the world if the water is clear : chlorophyll is at work
This is a clear proof that most shallow bottoms are more or less green at Caicos Bank, a fact well known to sea turtles and dugongs which feed there on turtlegrass
Try various values for RLBgb while modeling, and look at imageZ
- RLBgb=1.000 : this is the default "neutral" Soils Line
- RLBgb=1.300 : this is as high as I've ever seen

Calibration 5 : mask_4 shallow BDH
BiDimHistogram shows only marine pixels
which are located under mask_4

Did I hear you claiming that you lack the experience for interpreting such a diagram? I find it hard too!
Lsblue<Lswblue : did you notice some shallow pixels exhibit a slightly negative bottom contrast?

<<==============RLBgb1.22 ========

this states that, when modeling by the 2-bands case, the "SoilLine" is now a straight line that runs through LsM with a slope=1.22 in normalized data

Customizing the two-bands case : profileAB

About the Soil Line in the Blue vs Green 2-band case
-Z/MSL0.00/RLBgb1.000/cZ1.000
Syntax_Z
Syntax_ProfileAB
Profiles 1, 2, 3 and 4

White profile_1.shp
Green profile_2.shp
Blue profile_3.shp
Red profile_4.shp

nice -20 ./4SM.4.08 -Process/14_Aug_2011/05_AUG_2011 -OriginFUGRO-NPA @@
-DB/tmnov/31_1_0_0/4_1/1265_1559/208.365_2398.755/1_1                 @@
-Mis/Bahamas/Caicos/Landsat/TM/FUGRO-NPA/UTM_19_008/0.030_0.030/22_NOV_1990         @@
-LS/0255.0/255.0/255.0/255.0             @@
-cWL/0.500/0.602/0.500/0.500_AutoCAL         @@
-M/@000001/00002/00003/00004             @@
@CP/0000.0/000.0/000.0/000.0_1.52m         @@
-LsM/199.8/104.0/121.2/099.1_AutoCAL         @@
-LsM/199.8/108.0/121.2/099.1             @@
-LsM/219.2/120.0/135.6/111.1_cLM=1.130     @@
-Lsw/061.0/017.7/011.2/006.9_AutoCAL       @@
-Lsw/60.42/17.41/10.79/06.50                 @@
-dLsw000.0/000.0/000.0/000.0                @@
-Lw/0009.8/001.5/000.0/000.0_AutoCAL         @@
-Lw/0010.6/001.5/000.0/000.0             @@
-Lm/0001.0/001.0/001.0/255.0             @@
-KK1_2_0.5195/Knir4.500/mask_3_AutoCAL         @@
-Z/MSL0.00/n_1/RLBgb1.000/mask_3_AutoCAL    @@
@Z/MSL0.00/n_1/RLBgb1.220/mask_3         @@
-B/tclNe/Bmin0/cLM1 .13             @@
-deglint/vRbaD/GlintM25.0_AutoCAL     @@
-deglint/vRbaD/GlintM10.0         @@
@deglint/vRbaD/GlintM00.0         @@
-extract/v/fullBDH/NIRmax200.0/mBPL4/mask_4           @@
-calibrate/v/BDSNpZg/BDh_12_13_34         @@
-Model/mask_3                 @@
-Smooth/5/D/Smart+             @@
-ProfileAB/R/profile_1/ZM/Reg_2_3/0_201/0_30_2.0/*        @@
@ProfileAB/R/profile_2/ZM/Reg_2_3/0_201/0_30_2.0/*         @@
@ProfileAB/R/profile_3/ZM/Reg_2_3/0_201/0_30_2.0/*         @@
@ProfileAB/R/profile_4/ZM/Reg_2_3/0_201/0_30_2.0/*

Profile_1 RLBgb1.00	Profile_1 RLBgb1.00	Profile_1 RLBgb1.22	Profile_1 RLBgb1.22
Profile_2 RLBgb1.00	Profile_2 RLBgb1.00	Profile_2 RLBgb1.22	Profile_2 RLBgb1.22
Profile_3 RLBgb1.00	Profile_3 RLBgb1.00	Profile_3 RLBgb1.22	Profile_3 RLBgb1.22
Profile_4 RLBgb1.00	Profile_4 RLBgb1.00	Profile_4 RLBgb1.09	Profile_4 RLBgb1.09

Profiles 1, 2 and 3 require -Z/MSL0.00/n_1/RLBgb1.22
Profiles 4...............require -Z/MSL0.00/n_1/RLBgb1.09

We are learning quite something here

The lagoon ....bottoms are much brighter and very green : RLBgb1.22
The outer reef bottoms are much darker and less green : RLBgb1.09
The RLBgb trick does quite well indeed !!!!!!!!
- Provided we can customize the commandline to process the mask_4 area separately

Profile 6

toto

All the above is confused, but we are learning a lot.
Now let's work with one only set of profiles : profile_6

nice -20 ./4SM.4.08 -Process/14_Aug_2011/05_AUG_2011 -OriginFUGRO-NPA @@
-DB/tmnov/31_1_0_0/4_1/1265_1559/208.365_2398.755/1_1                @@
-Mis/Bahamas/Caicos/Landsat/TM/FUGRO-NPA/UTM_19_008/0.030_0.030/22_NOV_1990        @@
-LS/0255.0/255.0/255.0/255.0            @@
-cWL/0.500/0.602/0.500/0.500_AutoCAL        @@
-M/@000001/00002/00003/00004            @@
-LsM/219.2/120.0/135.6/111.1_cLM=1.130    @@
-Lsw/60.42/17.41/10.79/06.50                @@
-dLsw000.0/000.8/000.0/000.0    @@
@dLsw000.0/0000/000.0/000.0    @@
-Lw/0010.6/001.5/000.0/000.0            @@
-Lm/0001.0/001.0/002.0/255.0    @@
@Lm/0001.0/002.0/002.0/255.0    @@
-KK1_2_0.5195/Knir4.500/mask_3_AutoCAL        @@
-Z/MSL0.00/N_1/RLBgb1.220/mask_4            @@
-Z/MSL0.00/N_1/RLBgb1.090/mask_4            @@
@Z/MSL0.00/RLN_0.96_21.5/n_1/RLBgb1.220/mask_4 @@
@Z/MSL0.00/RLN_0.96_23.0/n_1/RLBgb1.220/mask_4 @@
-B/tclNe/Bmin0/cLM1 .13            @@
-deglint/vRbaD/mDEGLINT19/GlintM255.0    @@
-Model                 @@
-Smooth/5/D/Smart+            @@
@RegressZZv/p/seatruth*depth_points_reproject_pruned/ZM/2_3/0_201/0_10_1.0/*    @@
-ProfileAB/R/profile_6/ZM/Reg_2_3/0_201/0_30_2.0/*         @@

RLBgb=1.090

A vast majority of pixels are happy with RLBgb=1.090 with the two-bands case
- they are not so green after all
Only the darkest pixels under mask_4 over the outer reef yield badly overestimated depth with the two-bands case
- they are even less green than assumed by RLBgb=1.09
A minority of pixels would require RLBgb=1.220 with the two-bands case, as shown in the profile to the right

RLBgb=1.090

A vast majority of pixels are happy with RLBgb=1.090
- they are not so green after all
profile 6-1 from 0 to ~28 km :
- this section is happy with RLBgb=1.090
profile 6-1 from ~28 to ~50 km :
- as it would require RLBgb=1.220 with the two-bands case
- to account for much greener bottoms in the NW area

Seatruth 0-10 m

-RegressZZv/p/seatruth*depth_p....pruned/ZM/2_3/0_201/0_10_1.0/*

Seatruth 0-30 m
-RegressZZv/p/seatruth*....pruned/ZM/2_3/0_201/0_30_1.0/*

We are learning quite a lot here!
although this is Landsat TM imagery, 30 m pixel size.

Next we do seatruth for good: Calibration 6!