[sane-devel] How to get infrared image data with sane on an epson v700 scanner?

[m. allan noah]

On Tue, Feb 1, 2011 at 10:20 PM, Alain Culos <sane-devel at asoundmove.net> wrote:
> Thanks again for your help Olaf.
>
> For Allan's information, I'm still stuck in a mode where my posts don't show
> up on the list, so I'm relying on one of you to reply for them to appear on
> the list.
>

I will contact the alioth admins about it. Perhaps they dont like
something about your mailserver.

I'm leaving the rest of the email for others to see, but I will make
one comment- you will probably have to dig deep on this one, I don't
think anyone ever got this working before.

> I have managed to get scanimage to produce some results with IR mode from my
> scanner thanks to your help Olaf. However I seem to have a few problems with
> the data: format and contents.
>
> Scanimage seems to work normally with plain regular reflective scans
> (paper). So that's a good start.
>
>
> 1/ Format issues:
>
> First it seems that the tiff or the pnm produced by scanimage is interpreted
> strangely: it shows as if it was three frames, side by side, of B&W data.
>
> scanimage -p -d epson2 --resolution 150 --mode Infrared --depth 8 --format
> tiff -v --batch-count=1 --batch=IR-test.tiff
>
> scanimage -p -d epson2 --resolution 150 --mode Infrared --depth 8 --format
> pnm -v --batch-count=1 --batch=IR-test.pnm
> (scans 1272x1754)
>
> scanimage -p -d epson2 --resolution 150 --source "Transparency Unit" --mode
> Infrared --depth 8 --format tiff -v --batch-count=1 --batch=IR-test.tiff
>
> scanimage -p -d epson2 --resolution 150 --source "Transparency Unit" --mode
> Infrared --depth 8 --format pnm -v --batch-count=1 --batch=IR-test.pnm
> (scans 880x1453)
>
>
> Imagemagick complains thus (different numbers whether I go with TPU or not,
> but same errors):
>
>> convert IR-test.tiff IR-test.jpg
> convert: IR-test.tiff: Bogus "StripByteCounts" field, ignoring and
> calculating from imagelength. `TIFFReadDirectory' @ tiff.c/TIFFWarnings/703.
> convert: IR-test.tiff: Read error at scanline 584; got 2544 bytes, expected
> 7632. `TIFFFillStrip' @ tiff.c/TIFFErrors/493.
>
>> convert IR-test.pnm IR-test.jpg
> convert: unable to read image data `IR-test.pnm' @ pnm.c/ReadPNMImage/923.
> convert: missing an image filename `IR-test.jpg' @
> convert.c/ConvertImageCommand/2838.
>
>
> Like I say it shows as three frames side by side (left/right) - as if it was
> RGB data displayed as three B&W images: each channel as a full separate
> image. Only I expected the image to be a single grayscale channel.
>
>
> 2/ IR content issues:
>
> It seems like the data itself is not proper IR data. Without TPU, it looks
> like plain light, but I can't tell whether it is RGB or just grayscale. Ok,
> maybe that's normal, since I would expect the IR to only be active with the
> TPU.
>
> With TPU it looks like Grayscale data split in three separate images. It
> certainly does not look like IR data (I expect dark flecks of dust and only
> very slight shades of light gray based on the actual RGB pixel colour for
> the slides).
>
>
> Any way you would advise me getting my teeth into this. Debug settings, dump
> files to produce and look at. What part of the code should I start with
> rather than try and make sense of it all in one go?
>
>
> Thanks again for any hints.
>
> Regards,
> Alain.
>



-- 
"The truth is an offense, but not a sin"