go to Landsat 8 study cases

15 m GSD pansharpened OLIP for SDB
 

PAN band for shallow water work two peer-reviewed papers 2017

From DN to reflectance : made easy

OLI offsets : we deserve better

Noise : PANCHRO should be less noisy than MULTI

PAN-MULTI co-registration : must be improved for PAN sharpening

PAN response curve : is OK on the Blue side , should include NIR

Offsets: we deserve better!!
This is scene LC80120442013346LGN00 over Andros Island, Bahamas
The geometric consistency of the image is a must for shallow water work.
Clouds and haze at various altitudes are a world of complexity.
Haze MUST be deglinted.
This works reasonably well
if/provided the glint correlations do not include clouds.
Now, this deglinting is severely limited
by the geometric inconsistencies in the image (which pixels?).
see also Waimanalo WV2

FCC there is a ~6 pixels offset this is consistent accross the whole VISNIR bands but NOT with the SWIR1 band but NOT with the PAN band	SWIR1 in the SWIR1 image, the offset increases from ~6 to ~14 pixels, mailny in Y direction this is enough to explain the poor glint correlations of SWIR1 with VISNIR bands SWIR1 is not suitable for deglinting, unfortunately
This in itself is a cause of concern!	Now, this explains some weird observations!
As for the 2*2 sum-aggregated Panchromatic band, there is only a very small offset. This is enough to explain the poor glint correlations of the PAN band with other VISNIR bands.

Why is the OLI  PANCHRO band noisier than MULTI bands?