trws-deglint-2

**Deglint and investigate your image**
step 1: update your command line
step2: write a deglinted image
step 3: the dark pixel assumption step 4: run the Deglinter
step 5: investigate your image Run a profile Scatter plots for the red polygon Scatter plots for the green polygon Play time Run a profile
step 6: conclusion
next to Extract calibration data

Activity: use LINE_2

Smoothed and Deglinted image

Extraction window
for a comparison -and some fun-,
these illustrations show raw_data inside a -E70/610/150/320 subwindow.

In order to produce the above image

First run the script as it is with -Deglint/vRbaD/GlintM13
- with a @E70/610/150/320 subwindow disabled
- so that the whole image is deglinted (backdrop)
Then run the script with
- with -Deglint/vRbaD/GlintM0.0
  - so that no deglinting is performed at all: GlintMax=0
- with -E70/610/150/320 subwindow enabled,
  - so that only a sub-window is processed

Near Infra Red and Linearity

This Tarawa_XS image is the best image I know which demonstrates clearly,
- particularly in the NIR band,
- that shallow water optics is very much usable and useful;
- see another clear case for a TM image.
This is because of the extremely gently sloping of the sea bottom
- over a very clean, fine-grained and very bright carbonate mud:
- please google tarawa mud flat.
Just 15 cm of water is enough to attenuate the BOA NIR radiance
- to about half its reference value at null depth:
- 200/exp(2K*Z)=200/exp(4.5*0.15)=101
Claims that shallow water optics -as rendered through radiance measurements in remote sensing imagery- might not be used for water column correction get severely beaten!
- we shall see for ourselves that everything looks very much linear indeed after "linearization".
- therefore multiple reflection, inelastic scattering -and the like- do not appear to scramble things beyond reasonable for our operational application: the simplified RTE seems to be applicable.

Activity : use LINE_3

LINE_3

LINE 3: Run the Profiler

How to use the glint regressions to choose spectral Lsw
A windy scene
Deglinting may not generate radiances Ls<La!
Syntax_Lsw

profile_1 profile_2	Over open waters: the water surface exhibits strong wind/swell-modulated sea-surface glint Inside the lagoon: the water surface is not affected by swell/wave-modulated glint Cloud shadow: from South to North, the profile runs through cloud shadows at km 5 and 16 LswNIR was estimated ~=6.5 under the cloud shadow but we know that the radiance of darkest shadow is at 6 this is the path radiance LaNIR
profile_1 LswNIR=6.5 -dLsw000/000/000 try with -smooth/5/Smart+	profile_1 LswNIR=5.5 -dLsw0000/0000/0-1.0 try with -smooth/5/Smart+
Note the 8-bits quantization noise See at km 13 that radiances under the cloud shadow are too low in all three bands Lsw values must be decreased: Change for -dLsw000/0000/0-1.0	By application of -dLsw000/0000/0-1.0 Lsw is now adjusted to Lswgreen=25.9 Lswred=13.9 LswNIR=5.5 ?After deglinting, all three bands now exhibit approx atmospheric path radiance even over the darkest shadow

From now on: -dLsw/000/000/-1.0

back to Tarawa tutorial
back to Prepare your image