Optical calibration, bathymetry, water column correction and bottom typing of shallow marine areas, using passive remote sensing imageries

Full CASI mosaic at Mahone Bay, 2001
image courtesy of HDI
14350*10140 17-bands CASI image, pixel size 1 metre : 4SM license cost is 4974 Euros,   processed january 2006

WL/442.5/464.6/490.5/510.1/520.3/529.7/540.0/555.0/565.3/574.7/585.1/595.5/605.0/624.8/654.3/684.8/850.3 in nanometers

In spite of radiometric quality problems which are poorly controlled and still under investigation,

to show that 4SM is capable of processing a VERY VERY large image

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
I keep digging
until suitable data
become available
raw image

equalized image


computed depth


TCC of "normalized" spectral bottom reflectance



Bottom typing at Mahone Bay,
as of 04 septembre 2005

Image balancing has been slightly improved
(still not satisfactory though)

it is presented as a true color composite
of deglinted image, histeq enhancement

 Spectral Water-Column-Corrected
Shallow Bottom Reflectance

it is presented as a true color composite,
histeq enhancement
Spectral Shallow Bottom Reflectance

it is composited into a 3 bands BGR true color composite
through a process that optimizes the visual aspect
 Shallow bottoms are classified by 4SM
into 7 bottom types
and less than 1% unclassified




Optical Calibration Diagrams : 
a humbling experience for a hardy practitioner

One of the things I have learned lately: the BPL must be made 
to be parallel to the observed BPL pixels, rather than to match it closely: 
this is because the shallow bottom at BPL pixels is actually far from being spectrally neutral,
as it is now well known that "supposedly clean sands" actually harbour a host of "cyanobacterial biofilm" material
which is quite alive and active (this particularly applies to Tracady Bay dataset, as I recall).

This results is the Optical Water Type OIII+0.2 here,
which really belongs toJerlov's familly of curves of attenuation coefficients at all visible wavelengths,
as demonstrated in the calibration diagrams below.

It is quite clear that an orderly set of operational K values has been established,
although some further and minor refinements are still in order.













Calibration of bands_16_14_10_3

Bandset for MAHONE BAY 2001==============================
# 1 [16U] 420.5nm+/- 10.2nm (rows 272-282) DNSRU:1000.000000    NOT USED
# 2 [16U] 442.5nm+/- 10.3nm (rows 260-270) DNSRU:1000.000000
# 3 [16U] 464.6nm+/- 10.3nm (rows 248-258) DNSRU:1000.000000
# 4 [16U] 490.5nm+/- 10.4nm (rows 234-244) DNSRU:1000.000000
# 5 [16U] 510.1nm+/-  5.8nm (rows 226-231) DNSRU:1000.000000
# 6 [16U] 520.3nm+/-  4.8nm (rows 221-225) DNSRU:1000.000000
# 7 [16U] 529.7nm+/-  4.8nm (rows 216-220) DNSRU:1000.000000
# 8 [16U] 540.0nm+/-  5.8nm (rows 210-215) DNSRU:1000.000000
# 9 [16U] 555.0nm+/-  5.8nm (rows 202-207) DNSRU:1000.000000
#10 [16U] 565.3nm+/-  4.9nm (rows 197-201) DNSRU:1000.000000
#11 [16U] 574.7nm+/-  4.9nm (rows 192-196) DNSRU:1000.000000
#12 [16U] 585.1nm+/-  5.8nm (rows 186-191) DNSRU:1000.000000
#13 [16U] 595.5nm+/-  4.9nm (rows 181-185) DNSRU:1000.000000
#14 [16U] 605.0nm+/-  4.9nm (rows 176-180) DNSRU:1000.000000
#15 [16U] 624.8nm+/-  5.8nm (rows 165-170) DNSRU:1000.000000
#16 [16U] 654.3nm+/-  4.9nm (rows 150-154) DNSRU:1000.000000
#17 [16U] 684.8nm+/-  4.9nm (rows 134-138) DNSRU:1000.000000
#18 [16U] 850.3nm+/- 25.0nm (rows  37- 62) DNSRU:1000.000000
#19 [16U] 922.0nm+/- 22.1nm (rows   1- 23) DNSRU:1000.000000    NOT USED