A 1630*1860 SPOT-1 image of Tarawa atoll, Kiribati, 1986
4SM license key for this image is 203 Euros.
return to Kiribati
Over 30,000 inhabitants on this 30 km wide atoll. A fierce stronghold of the Japanese Navy during WW2.
A stuffy atmosphere under the Equator, strong wind, lots of clouds, skyglint everywhere.

Deeper than 4-6 m, depth is computed using XS1 band only,
and assuming a bright and uniform bottom substrate.

Please notice that, deeper than 4-6 m in this water type, a Blue band with K[blue]~=K[green] would not help :
this lagoon should make for a good case of combined use of a SPOT panchromatic channel
in order to allow modeling in the 4-6 to 20 m depth range.

Please refer to tutorial dataset for more details


root enhancement
A pretty nice image...

histeq enhancement
... not that good anymore, hey!


Carefull deglinting is absolutely necessary
before this image may be used.

The One-Band case 
allows to turn out a usefull depth map deeper than 4-6 m

Maximum depths of ~16 m just north of the main pass
are confirmed by nautical chart. Poor results at some locations, though:  the main clouded areas, and all cloud shadows as well,  should better be masked out.
The One-Band case 
is signaledby the choice of a constant bottom reflectance deeper than 4-6 m.

Foul waters in the vicinity of densely inhabited areas
make it necessary to raise progressive Lm thresholding to fairly high values in order to prevent heavy "dark-bottom" artifact from invading the results".

 Carefull deglinting, followed by smart-smoothing, ...
 ...then progressive Lm thresholds, together with the one-band case deeper than 4-6 m, have done quite a good job.

In-flow of cristal-clear oceanic water through the reef rim cause an "ultra-bright" bottom reflectance artifact ...
... without any noticeable artifact in the depth image though.

Calibration diagram
Jerlov's water type OIII+0.12 for the clearest waters in this lagoon. Maximum depth penetration of XS1 channel if of ~20 m over bright bottoms.
Please notice that , in this water type, a Blue band with K[blue]~=K[green] might not help deeper than 4-6 m.

In-flow of very clear ocean waters through the reef rim at some specific locations is the cause of a distinct bulge of the Brightest Pixels Line at very shallow depths: see how this results in "deep blue ultra-bright" bottom substrates in image_B: this of course is a local artifact.



Créer un site
Créer un site