A time series of Landsat 8 OLIP images
at Nosy Be, Madagascar

work done in may 2020 at 15 m GSD
Using the Panchromatic band for water column correction
to derive water depth and spectral bottom signature:

Landsat 8 OLIP bandset used for this work

Purple=1Blue=2Green=3PAN=4Red=5NIR=6 and SWIR1=7

   peer-reviewed and published in 2017    

March 27th 2018
processed in a 32-bits Ubuntu 16 LTS install,
on a VirtualBox for Windows, on my new laptop

OpenEV now available!
All the above was done before I could use OpenEV
Channel_6Special effects : mask mSE

Channel_7 : Green and Red solutions
Channel_9 : Z in decimeters

Channel_11 : Average bottom brightness B chequered
Until the image is licensed,
4SM writes out chequered results
Cells are 500*500 pixels.
The license cost for this im ageis 758_Euros

Channel_35 : Z in centimeters

Average bottom brightness B

  • Circa Row_3300 and Line_1900, in this all-terrigeneous environment (any submerged vegetation?) , Yellow_Brown hues 
    • smack of more turbid waters (like if waters flowing Southwards),
    • and therefore entails some depth underestimation
  • South of Line_3100 : Reddish hues
    • indicate areas where optical calibration applies comfortably

TOA TCC deglinted


View as if at "low tide". Shallow substrates are
much darker than soils on land.   Lmred=1.
Water turbidity along the water lineshows up clearly;  this causes depth  underestimation.

TCC raw 16U data
Waters there might be less clear.

Darker patches: is there any upwelling?

Water turbidity causes
  • depth underestimation by the Red solution
  • optically deep waters much closer to the waterline

Channels_41_39_3 : TCC raw 16U data
What's this fabric in the raw data?
This is a very bad noise,
heavy smoothing required.

Channel_42 : NIR band
Why is the NIR signal smoothed like that?

Profile Blue
Raw radiances are very noisy.
This noise is NOT correlated among spectral bands,
therefore it is not caused by specular reflection.
This is appalling

Raw DN 12-bits radiances,
on a scale of 0-4095, are very noisy,
  • noise ranges over ~15 DNs,
  • even exhibiting "worm-like features" which remind me of fancy wavelet transforms or the like.
  • Panchro in channel_3 is even more noisy, ranging over ~40 DNs, although it is over-sampled from 0.5 to 2 m GSD.
Or did I miss something?
  Atmospheric adjacency effect
  • This is consistent with a fairly strong atmospheric  adjacency effect which needs absolutely to be "de-glinted".
  • "Deglinting" the atmospheric adjacency effect, like if it was mere specular reflection at the sea surface, actually works very well
    • after all, it is caused by the multiple diffusion of the land-reflected photons away from the direct field of view of the sensor telescope
    • this diffusion is caused by the abundance of aerosols
Atmospheric path radiance
  • For a NIR band to score 150 over a range of 0-4096, it requires quite a lot of aerosols.
  • For a NIR band to score 34 over a range of 0-255 (where "bright" Tuscany terrigeneous beaches score ~255), it requires quite a lot of aerosols in august.
  • This reminds me of Abu Dhabi overcast skies.
  • This is consistent with a fairly strong adjacency effect which needs absolutely to be "de-glinted".
Massimo:  Heavy smoothing of costly 2 m GSD data does not make sense to me.
Why not try Landsat 8 data, pansharpened to 15 m GSD, see if it can save you money for your project of monitoring Tuscany's coastal erosion?
Note:  the PLEIADES Panchromatic band should be discarded altogether
#acquired geotiff 16 bit (not JPEG2000)
#Line_0 - create the database costa_toscana.pix and run the AutoCalibrator
nice -20 ./4SM.8.06 -Process/23_Mar_2018 -Origin/MassimoPerna                                             @@
-DB/costa_toscana_sub/34_3S_5S_0s/5_38/7054_6416/615.975_4794.925                                         @@
-Mis/Italy/Costa_toscana/PLEIADESP/1A/NA/bOA/UTM_32_008/0.002_0.002/03_AUG_2017        @@
-MakePIX/ISA                                                                                 @@
-Import/v/dTM/data*costa_toscana_sub/dbnc_0_0S_5S_0s/R7054_L6416/Origin_615.975_4794.925_0_0/chIn1,5/chOut38,42 @@
-SCL/0070.0/0065.0/0075.0/0060.0/0110.0_SCL                                                            @@
-LS/0010000/10000/10000/10000/10000                                                               @@
-KK1_2_0.50/Knir4.5/CoefK_1.0/mask_3                                                                @@
-deglint/vrbaD/GlintM50/Ext_3300_3600_2800_3200                                                        @@
-extract/v/RawBDH/FullBDH/NIRband5/NIRmax5/mBPL2    /mask_3                                                @@
-calibrate/voAi/BdSNpzg/BC_5_4_3_2/BDh_12_13_34                                                        @@

#Line_1 - Run the Modeler............Run the Modeler............Run the Modeler............
nice -20  ./4SM.8.06 -Process/23_Mar_2018  -Origin/MassimoPerna                    @@
-DB/costa_toscana_sub/34_3S_5S_0s/5_38/7054_6416/615.975_4794.925                @@
-SCL/070.0/065.0/075.0/060.0/110.0_SCL                                @@
-LS/010000/10000/10000/10000/10000                                @@
-Lsw/196.5/150.2/097.0/089.4/034.6_Lsw_AutoCal                            @@
-dLsw/-2.5/0-2.0/003.0/006.0/000.0                                @@
-Lw/0027.0/015.0/000.0/000.0/000.0_Lw_AutoCal                            @@
-M/@000001/00002/@0003/@0004/00005_M                                @@
-Lm/0001.0/001.0/001.0/005.0/255.0_Lm                                @@
-LsM/241.0/218.7/208.2/231.2/130.4_LsM_cLM=0.900                        @@
-CP/219.80/175.4/000.0/098.1/000.0_5.00m                            @@
-KK1_2_0.63/Knir4.5/CoefK_0.0/mask_3                                @@
-Z/NA0.00d/N_1/nDR_37_1000_zDTM/land_8/mask_3                            @@
-B/tclNe/cNorm2.50/Bmin0/cLM 2                                     @@
-deglint/vrbaD/GlintM0 50/Ext_3300_3600_2800_3200                            @@
@Extract/v/rawBDH/FullBDH/NIRband5/NIRmax5.0/mBPL2/Ext_0001_7054_0010_0400            @@
@Extract/v/rawBDH/FullBDH/NIRband5/NIRmax5.0/mBPL2/Ext_0001_7054_0800_1200            @@
@Extract/v/rawBDH/FullBDH/NIRband5/NIRmax5.0/mBPL2/Ext_0001_7054_1800_2200            @@
@Extract/v/rawBDH/FullBDH/NIRband5/NIRmax5.0/mBPL2/Ext_0001_7054_2800_3200            @@
@Extract/v/rawBDH/FullBDH/NIRband5/NIRmax5.0/mBPL2/Ext_0001_7054_3800_4200            @@
@extract/v/rawBDH/FullBDH/NIRband5/NIRmax5.0/mBPL2/Ext_0001_7054_4800_5200            @@

-extract/v/rawBDH/FullBDH/NIRband5/NIRmax5.0/mBPL2 /Ext_0001_7054_5800_6200            @@
-calibrate/voAI/BdSNpzg/BC_5_4_2_1/BDh_12_23_34                            @@
-Model/v/Rt1.10/dTM/ZM30/mask_3                                    @@
-Smooth/5/D/smart+    -LL10                                    @@
-TpsV/1/3/Bmin150/RLD4/CK1.001        @-Eshp      @-E2shp    @-DDshp                @@
@ProfileAB/lZrvWBc/profile_black/LsMax_400/chAB_1_2_4                               @@
@ProfileAB/LzrvWBc/profile_red/LsMax_400/chAB_1_2_3_4_5                               @@
@ProfileAB/lZrvWBc/profile_red/LsMax_400/chAB_1_2_4                                   @@
@ProfileAB/LzrvWbc/profile_blue/LsMax_600/chAB_1_2_3_4_5                              @@
@ProfileAB/lZrvWBc/profileAB_1900_2800_0100_0100/LsMax_400/chAB_1_2_4                          @@


Créer un site
Créer un site