Optical calibration, bathymetry, water column correction and bottom typing of shallow marine areas, using passive remote sensing imageries
 
Two  4-bands IKONOS images of Kauai, Hawaii Islands
courtesy of Ron Abileah
papers by Abileah
papers by Abileah
US patent by Abileah

Wavelengths at mid-response curve are 480, 551, 665 and 805 nm

These images also have a Panchro band at 2 m resolution
This dataset is covered by SHOALS LIDAR bathymetry,
and CASI hyperspectral data

Work done on Sept 13-14th 2011

 


 
1 - NO NEED for field data, nor for atmospheric correction
2 - this is demonstrated in this website, using a variety of hyper/multi spectral data
 
Requirements are
1 - homogeneous water body and atmosphere
2 - some coverage of optically deep water
3 - some coverage of dry land
 
Problems are
1 - the precision on estimated depth is found wanting, because the noise-equivalent change in radiance  of accessible data is too high for shallow water column correction work 
2 - radiance data should be preprocessed by the provider at level 1 in order to improve S/N ratio
3 - exponential decay: the deeper/darker the bottom, the poorer the performances
 
So
I keep digging
until suitable data
become available
 

As of january 2012, it turns out that this IKONOS data,
and the Oahu WV2 data as well, 
were "offset" by an unknown amount:
"the band images after subtracting an estimate of the deep water upwelling radiance".
This interferes with my deglinting process:
I need the original radiance, direct from data provider


return to 4SM study cases
po 180000 0000000
po 180000 0010000


po_180000_0000000

2 - kauai0.zip==================
This is my working directory for po_180000_0000000
PCIDSK database is kauai0.pix see database  structure
Use as is, or export to your prefered image format
Glint in the NIR band is very poorly correlated to other three bands
No beach
Raging breaking waves
Clouds
Had to use another 4SM trick : now got a quite handsome results
Water there appears to be the clearest oceanic type
Did not use those 400000 plus SHOALS depth points...
Could not estimate green wavelength other than at mid-waveband;
           therefore the slope might be important in FinalZ = Tideheight + slope * Computed depth
See kauai0.pix
Depth in centimeters is in channel_32 : quite interesting
Water column corrected TCC is in channels_30_29_28 RGB
Deglinted TCC is in channels_26_25_24
Results are noisy as a result of choppy seas and sturdy glint




True Color Composite : Why that blocky fabric?
I find it hard that this is the original data direct from  DigitalGlobe
True Color Composite : Why that blocky fabric???

 
Raw image, true color composite
is in channels_3_2_1
 
Raw image, true color composite: zoom
is in channels_26_25_24
 
image_Z250
is in channel_8
image_ZG in centimeters
is in channel_32
 
 
image B
is in channel_10
Source Image ID: 2004042821282980000011610601
Product Image ID: 000
Sensor: IKONOS-2
Acquired Nominal GSD
Cross Scan: 0.82 meters
Along Scan: 0.82 meters
Scan Azimuth: 179.98 degrees
Scan Direction: Reverse
Panchromatic TDI Mode: 13
Nominal Collection Azimuth: 146.6109 degrees
Nominal Collection Elevation: 84.38902 degrees
Sun Angle Azimuth: 111.7189 degrees
Sun Angle Elevation: 72.15586 degrees
Acquisition Date/Time: 2004-04-28 21:28 GMT
Percent Cloud Cover: 27
image LBS TCC
is in channels_30_29_28
  • this is a "low-tide", or "water column corrected", RGB true color composite of the shallow bottom reflectance over the scene (linked to a false color composite)
  • in this view, each channel is scaled linearly over the range 0-250 for both dryland and shallow areas:
    • 0     for the BOA reflectance of a black body
    • 250 for the BOA reflectance of the brightest shallow bottom type that exists over the scene
Calibration

Not for a beginner!
Hard to see through,
as there may not be any bright sand around

Oceanic I type of water is a sensible statement.

Impossible to investigate
the question of wavelength for the green band : wavelengths at mid wavebands for all four bands

NO FIELD DATA USED YET

BiDimensional histograms
and regressions
over open waters
 
 Here it can be seen that the glint correlations
are lousy, to say the least
skyglint, sunglint, and white caps :
for interesting clues, please refer to US patent by Abileah



 
 

Kauai USACE LIDAR coverage

Kauai image coverage
 

Maui USACE LIDAR coverage
 

Oahu USACE LIDAR coverage
 

Modify Channel Descriptor file="kauai0.pix";

cm01=" 1 1_Scl'd_inrange=0,180 480 nm"; dboc=1 ; r mcd
cm01=" 2 2_Scl'd_inrange=0,180 551 nm"; dboc=2 ; r mcd
cm01=" 3 3_Scl'd_inrange=0,180 665 nm"; dboc=3 ; r mcd
cm01=" 4 4_Scl'd_inrange=0,160 805 nm"; dboc=4 ; r mcd
cm01=" 5 Special_Effect_mask; use pctmSE"; dboc=5 ; r mcd
cm01=" 6 wZ: use pctwZ"; dboc=6 ; r mcd
cm01=" 7 wR: use pctwR"; dboc=7 ; r mcd
cm01=" 8 Z0->250 in dm: use pctZ250"; dboc=8 ; r mcd
cm01=" 9 Z: use pctZ20"; dboc=9 ; r mcd
cm01="10 B: use pctB"; dboc=10 ; r mcd
cm01="11 Bottom Typing: use pctClassified"; dboc=11 ; r mcd
cm01="12 LSBCC_blue "; dboc=12 ; r mcd
cm01="13 LSBCC_green"; dboc=13 ; r mcd
cm01="14 LSBCC_red "; dboc=14 ; r mcd
cm01="15 dZED: use pctZ250"; dboc=15 ; r mcd
cm01="16 BPL_12 "; dboc=16 ; r mcd
cm01="17 BPL_13 "; dboc=17 ; r mcd
cm01="18 BPL_23 "; dboc=18 ; r mcd
cm01="19 void"; dboc=19 ; r mcd
cm01="20 void"; dboc=20 ; r mcd
cm01="21 void"; dboc=21 ; r mcd
cm01="22 void"; dboc=22 ; r mcd
cm01="23 void"; dboc=23 ; r mcd

cm01="24 1_Ldeglinted_480 nm"; dboc=24 ; r mcd
cm01="25 2_Ldeglinted_551 nm"; dboc=25 ; r mcd
cm01="26 3_Ldeglinted_665 nm"; dboc=26 ; r mcd
cm01="27 4_Ldeglinted_805 nm"; dboc=27 ; r mcd

cm01="28 1_LSB_480 nm"; dboc=28 ; r mcd
cm01="29 2_LSB_551 nm"; dboc=29 ; r mcd
cm01="30 3_LSB_665 nm"; dboc=30 ; r mcd
cm01="31 4_LSB_805 nm"; dboc=31 ; r mcd

cm01="32_ZG: Zshallow in centimeters, others at -1"; dboc=32 ; r mcd

cm01="33_DTM: ZR in centimeters, others at -1"; dboc=33 ; r mcd

cm01="34 1_Raw_480 nm"; dboc=34 ; r mcd
cm01="35 2_Raw_551 nm"; dboc=35 ; r mcd
cm01="36 3_Raw_665 nm"; dboc=36 ; r mcd
cm01="37 4_Raw_805 nm"; dboc=37 ; r mcd

WLMin[ 1]=444.5 WL[ 1]=480.2 WLMax[ 1]=515.9
WLMin[ 2]=506.4 WL[ 2]=551.0 WLMax[ 2]=595.5
WLMin[ 3]=631.8 WL[ 3]=664.8 WLMax[ 3]=697.7
WLMin[ 4]=757.3 WL[ 4]=805.0 WLMax[ 4]=852.7
 
 
 

po_180000_0010000

Raw image, true color composite
is in channels_3_2_1
 
Deglinted image, true color composite
is in channels_26_25_24
 

Glint regressions
Source Image ID: 2003101921341100000011619176
Product Image ID: 001
Sensor: IKONOS-2
Acquired Nominal GSD
   Cross Scan: 0.90 meters
   Along Scan: 1.00 meters
Scan Azimuth: 180.02 degrees
Scan Direction: Reverse
Panchromatic TDI Mode: 13
Nominal Collection Azimuth: 351.1580 degrees
Nominal Collection Elevation: 63.99443 degrees
Sun Angle Azimuth: 157.7892 degrees
Sun Angle Elevation: 55.72259 degrees
Acquisition Date/Time: 2003-10-19 21:34 GMT
Percent Cloud Cover: 23
BiDimensional histograms
and regressions
 over open waters
 
BiDimensional histograms over open waters:
Can we trust the glint signal?
  • It can be seen that the radiometric quality of the signal captured by the sensor is questionable, to say the least
    • this raises serious questions
  • Using the NIR-based regressions  yields very poor results
  • Using the RED-based regressions improves matters significantly
The radiometric setup
of this image (LsM, Lsw, Lw)
is not under control:
Not worth the pain.
TCC
Very low radiances in shadow areas are questionable
FCC
Very low radiances in shadow areas are questionable

  • This is an enormous amount of glint : can't be removed to satisfaction

  • It turns out that this data was "offset" by an unknown amount: I need the original radiance, direct from data provider