water column correction of HICO images

28 narrow wavebands
BAND 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 WL 407 418 430 441 453 464 476 487 499 510 521 533 544 556 567 579 590 602 613 625 636 647 659 670 682 693 PAN 842

Image preparation
Optical calibration
home

HICO image preparation

I exported from OSU's ENVI format into PCIDSK format, using OpenEV
The image is in Level 1B, pixel size is 100 m, 500 columns, 2000 lines
I did not use the projection information provided
HICO is comprised of 87 raw bands, ~ 5 nm wide
presently, I can only manage 28 bands, in spite of my full HD screen display: see command line
So, I binned HICO raw bands as follows

WLMin[ 1]=401.2 WL[ 1]=406.9 WLMax[ 1]=412.7    by binning bands HICO_1 and HICO_2
WLMin[ 2]=412.7 WL[ 2]=418.4 WLMax[ 2]=424.1    by binning bands HICO_3 and HICO_4
WLMin[ 3]=424.1 WL[ 3]=429.9 WLMax[ 3]=435.6    ....
WLMin[ 4]=435.6 WL[ 4]=441.3 WLMax[ 4]=447.0    ....
WLMin[ 5]=447.1 WL[ 5]=452.8 WLMax[ 5]=458.5    ....
WLMin[ 6]=458.5 WL[ 6]=464.2 WLMax[ 6]=469.9    ....
WLMin[ 7]=470.0 WL[ 7]=475.7 WLMax[ 7]=481.4    ....
WLMin[ 8]=481.4 WL[ 8]=487.1 WLMax[ 8]=492.9    ....
WLMin[ 9]=492.9 WL[ 9]=498.6 WLMax[ 9]=504.3    ....
WLMin[10]=504.3 WL[10]=510.0 WLMax[10]=515.8    ....
WLMin[11]=515.8 WL[11]=521.5 WLMax[11]=527.2    ....
WLMin[12]=527.2 WL[12]=533.0 WLMax[12]=538.7    ....
WLMin[13]=538.7 WL[13]=544.4 WLMax[13]=550.1    ....
WLMin[14]=550.2 WL[14]=555.9 WLMax[14]=561.6    ....
WLMin[15]=561.6 WL[15]=567.3 WLMax[15]=573.0    ....
WLMin[16]=573.1 WL[16]=578.8 WLMax[16]=584.5    ....
WLMin[17]=584.5 WL[17]=590.2 WLMax[17]=596.0    ....
WLMin[18]=596.0 WL[18]=601.7 WLMax[18]=607.4    ....
WLMin[19]=607.4 WL[19]=613.2 WLMax[19]=618.9    ....
WLMin[20]=618.9 WL[20]=624.6 WLMax[20]=630.3    ....
WLMin[21]=630.4 WL[21]=636.1 WLMax[21]=641.8    ....
WLMin[22]=641.8 WL[22]=647.5 WLMax[22]=653.2    ....
WLMin[23]=653.3 WL[23]=659.0 WLMax[23]=664.7    ....
WLMin[24]=664.7 WL[24]=670.4 WLMax[24]=676.1    ....
WLMin[25]=676.2 WL[25]=681.9 WLMax[25]=687.6    ....
WLMin[26]=687.6 WL[26]=693.3 WLMax[26]=699.1   by binning bands HICO_51 and HICO_52
WLMin[27]=699.1 WL[27]=704.8 WLMax[27]=710.5   this is a synthetic Panchromatic band
WLMin[28]=836.6 WL[28]=842.3 WLMax[28]=848.0   by binning bands HICO_77 and HICO_78

The 4SM synthetic panchromatic band was prepared
by binning bands 4SM-7 to 4SM-25

WLMin[ 7]=470.0 WL[ 7]=475.7 WLMax[ 7]=481.4    ....
WLMin[ 8]=481.4 WL[ 8]=487.1 WLMax[ 8]=492.9    ....
WLMin[ 9]=492.9 WL[ 9]=498.6 WLMax[ 9]=504.3    ....
WLMin[10]=504.3 WL[10]=510.0 WLMax[10]=515.8    ....
WLMin[11]=515.8 WL[11]=521.5 WLMax[11]=527.2    ....
WLMin[12]=527.2 WL[12]=533.0 WLMax[12]=538.7    ....
WLMin[13]=538.7 WL[13]=544.4 WLMax[13]=550.1    ....
WLMin[14]=550.2 WL[14]=555.9 WLMax[14]=561.6    ....
WLMin[15]=561.6 WL[15]=567.3 WLMax[15]=573.0    ....
WLMin[16]=573.1 WL[16]=578.8 WLMax[16]=584.5    ....
WLMin[17]=584.5 WL[17]=590.2 WLMax[17]=596.0    ....
WLMin[18]=596.0 WL[18]=601.7 WLMax[18]=607.4    ....
WLMin[19]=607.4 WL[19]=613.2 WLMax[19]=618.9    ....
WLMin[20]=618.9 WL[20]=624.6 WLMax[20]=630.3    ....
WLMin[21]=630.4 WL[21]=636.1 WLMax[21]=641.8    ....
WLMin[22]=641.8 WL[22]=647.5 WLMax[22]=653.2    ....
WLMin[23]=653.3 WL[23]=659.0 WLMax[23]=664.7    ....
WLMin[24]=664.7 WL[24]=670.4 WLMax[24]=676.1    ....
WLMin[25]=676.2 WL[25]=681.9 WLMax[25]=687.6    ....

So, for 4SM, my database.pix 16U image database has 28 raw data bands, ~10 nm wide

Blue is band_7
Green is band_13
Red is band_22
Panchro is band_27
NIR is band_28

HICO images are not georeferenced.

Optical calibration

This by large benefited from investigations conducted on HYPERION at Fakarava atoll

HICO at Lee Stocking Island
June 2010 November 2010

**Results June versus November**
June 2010	November 2010
june no tide correction ARTIFACT: note the shoaling artifact over the very dark seagrass area	november no tide correction Note the absence of a shoaling artifact over the very dark seagrass area

june no tide correction	november no tide correction

These two LSI HICO scenes are not co-registered; so results may not be precisely compared.
As regards vegetation: one scene is taken in end spring, one in late autumn.
It takes very strict fine-tuning to obtain "reasonable" results.
I'm not allways sure what aspect of the "physics behind" such fiddling actually portends.
All this practise must be reviewed and investigated by competent accademic research.
But 4SM possibly is a viable alternative
to more complex semi-analytical approaches.