Support: 8/Dec/2008, Abdulwasey Mohammed, GB06/07

[mggr]

Initial contact from Abdulwasey, who was having problems processing atmospheric corrected versions of his data.

Project is GB06-07_241a_Silverflowe (Eagle, Hawk, ATM, though he may not be using Hawk?). See ticket #63 for processing/delivery info.

[mggr]

Some initial problems with azgcorr usage (Mark suggested some things).

Major problem appeared to be that he had an atmospherically corrected image file in bsq format and was trying to convert this to HDF to process it with azgcorr.

Sent him some suggestions on:

• using an external data file with azgcorr
• projection parameters for UTM/WGS84 zone 30
• azgcorr-DEM creation from LIDAR

[mggr]

Got an email and a call - apparently he doesn't have the original .hdf files(!), which would make it impossible to process anything. Tarballed up the eagle .hdr and .hdf files and sent those to him.

Need to send him an example command.

[mggr]

Linking processing/delivery ticket.

[mggr]

Seems he did have the HDFs :) His timeline is tight too - thesis submission in Feb!

Sent a set of example commands demonstrating how to use a bseq external file (created with ENVI).

If that works ok, the next step is to map the ATCORR output.  You said it was written as a bseq file - that's fine.  To use a bseq file for image input, you need to do something like this:

 azgcorr -mUTMZs 3 30 -1 lev1/e241a011b.hdf -3 lev3/test1-atcorr.hdf -bl 30 15 7 -1 -l 0 500 -Bs lev1/atcorr.bseq -B 4029 252 1 1 0 0
 azexhdf -h lev3/test1-atcorr.hdf -G lev3/test1-atcorr.tif 

Note that the -B parameters include a scale and offset for cramming the output into uint16. This is likely to cause confusion later - I warned him, but it may be good to re-emphasise later.

[mggr]

His output came out all yucky - looks like he's overwritten the original .bil files with atcorr bseq output, hence the mess. Offered some advice on how to proceed.

[mggr]

He was still having problems so uploaded one of his atcorr bseqs to the FTP server for us to look at.

The problem was the file was unsigned int, not float32 - the command I gave him was for float32 data.

Success with:

azgcorr -mUTMZs 3 30 -1 lev1/e241a011b.hdf -3 lev3/test1-atcorr.hdf -bl 30 15 7 -1 -Bs lev2-atcorr/a241a011b-bseq_atm66.bsq -B 4029 252 0 0 0 0 -l 0 500
azexhdf -h lev3/test1-atcorr.hdf -G lev3/test1-atcorr.tif
 (maps the first 500 lines of the atcorr output and creates a GeoTIFF for easy viewing)

Recommended bil output:

 azgcorr -mUTMZs 3 30 -1 lev1/e241a011b.hdf -3 lev3/test1-atcorr.hdf -Bs lev2-atcorr/a241a011b-bseq_atm66.bsq -B 4029 252 0 0 0 0 -BIout
 (no azexhdf line needed)

[mggr]

Sent Abdulwasey the -B parameters for all the other Eagle flightlines.

This is basically just -B no_of_lines no_of_bands 0 0 0 0. Really it should be read from the .hdr file ; Bill has said the external file stuff was added fairly quickly for AKW and isn't polished yet. Something for the next set of minor fixes, perhaps :)

[mggr]

Got a mail back from him saying it worked :)

He had an issue with a LIDAR DEM but managed to get a 10m DEM from elsewhere. Offered to fix up that DEM if he needs it.