Aerial Photograph of Kauehi atoll, French Polynesia

processed in 1998

 



 
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
 
So
I keep digging
until suitable data
become available
 

 

This color aerial photograph represents the area of the village of Kauehi,
Tuamotu archipelago, French Polynesia.

The photo was shot from an altitude of 2290 m at 8:35 am, at a scale of 1:1500. The camera's field of view is 45 degrees from vertical. The original negative AGFA AVIPHOT Color N 200 PE1 film was scanned at a 360 dpi resolution: 3251 lines and 2856 columns.

The color composite [ A3 ]  is enhanced through a logarithmic lookup table, where shades of blue are therefore linearly related to changes of water depth. With Lsw values of 32, 16 and 14, and bright sandy beach at approximately 255 in the Blue, Green and Red bands respectively, this 8-bits image exhibits an excellent dynamic range, ideal for shallow water modelling.

According to AGFA facts sheet, spectral sensivity bands are 380 to 480 nm for the blue band, 510 to 580 nm for the green band with a peak at ~560 nm , and 580 to 660 for the red band with a peak at ~ 640 nm.

Ratios of K[blue]/K[green]~=0.64, K[blue]/K[red]=0.303 and K[green][Kred]~=0.47 are observed. It is not possible to reconcile these observations with peak sensivity wavelengths reported above. Quite obviously, and probably because of their wideband character, these wavebands exhibit a Panchromatic behaviour in respect of attenuation processes, i.e. a progressive decrease of K values as the height of the water column increases.

At this stage, we do not know how to assign a "central wavelength" to these widebands. In particular, the Red waveband would appear to behave more like a "yellow" band, penetrating far deeper than a truly red band. Therefore, the callibration of attenuation coefficients by use of Jerlov's data is still uncertain at this early stage of the development of shallow water modelling to aerial photographs.

The results presented here were obtained using 2K values of 0.1336, 0.2077 and 0.4416 m-1 for the blue, green and red bands tespectively, with wavelengths "set" at 465, 583 and 593 nm. They are as an indication that, subject to research and development efforts into several specific aspects, shallow water modelling using color aerial photography has a potential of being developed into an operational procedure in a reasonable future.

DEPTH RESULTS [ A3 ]  ================================

Because a NIR band is not available, the waterline is not precisely determined. It is quite likely that all depths obtained should all be multiplied by a single correcting factor to be derived from ground truth. The results presented here were obtained through a 3x3 smoothing of the shallow water pixels prior to modelling.

BOTTOM REFLECTANCE RESULTS [ A3 ] ===================

It is not understood why the bottom reflectance should decrease markedly deeper than ~6-7 m.