sabahOLI20140205

February 5th 2014
scene LC81180562014036LGN00 Image courtesy of the U.S. Geological Survey

Data preparation
PAN sharpening using Rstudio

I realized that SWIR and NIR must not be sharpened,
as they do not exhibit significant light penetration

Sharpening the SWIR band
creates artifacts by transfering LowPassFilter shallow features from PAN to SWIR (and NIR).
Then, using SWIR (or NIR) for deglinting transfers this low-radiance noise to sharpened bands.
This creates artifacts which I noted in the Retrieved depth images at SanLorenzoChannel.

Sharpening the RED band
Note the LowPassFilter shallow features at very low radiance

Sharpening the GREEN band

Sharpening the COASTAL band

Conclusion

SWIR and NIR bands must not be sharpened
- they just need to be resampled from 30 to 15 m GSD
RED, GREEN, BLUE and COASTAL shall be sharpened

The data

TCC	TCC BOA deglinted
Glint regressions are excellent Calibration is limited to area under orange mask 3

Optical calibration
for the open waters area under orange mask_3

Optical calibration diagram for bands Blue, Green, Red and NIR	Optical calibration diagram for bands Coastal, PAN, Red and NIR
OII+0.3 water type of Jerlov

Ready for modeling
the whole scene

Retrieved depth see legend	TCC BOA deglinted see legend for image B

GSD 15 m Landsat 8 2014	GSD 30 m Landsat TM 1991 for a comparison

Optical calibration
for the open waters area under white ROI

Hydrology Location of profile_White and White coastal ROI Over the Blue range: deep water radiance is seen to decrease steadily from open waters towards the coast line this is because of "yellow substances" Other bands: are just slightly affected Altogether, water optical properties remain fairly constant: OII+0.3 water type of Jerlov But see profile_white for deep water radiances	Profile_white runs through fairly homogeneous waters type OII+0.3 of Jerlov, although the influence of the landmass is seen to increase steadily in the Blue bands as the profile gets closer to the water line Waypoint_3 is where deep water radiances is sampled for this calibration Waypoint_2 is where deep water radiances is sampled for the previous calibration It is seen that a Lsw is lower at Waypoint_3 than at Waypoint_2
Optical calibration diagram for bands Blue, Green, Red and NIR	Optical calibration diagram for bands Coastal, Green, Red and NIR

Second modeling along the coast
only pixels under red mask have been updated

Second modeling: Retrieved depth only pixels under Red mask have been updated	First modeling: Retrieved depth for a comparison

There is a good prospect that depths retrieved along the coast
are just as good as further offshore