[sane-devel] multipage scans

Carl Karsten carl at nextdayvideo.com
Mon Dec 14 21:19:39 UTC 2015


gscan2pdf - also neat.

I see it does deskew - perhaps you can give me some advice.

The rest is kinda OT for sane-dev. happy to move if someone can suggest a
more appropriate list.

I print this:

http://veyepar.nextdayvideo.com/main/recording_sheets/252/.pdf?day=12&room=OGGB_FP

As those talks are happening, write the start/end and maybe some notes.

I scan it:

http://veyepar.ndv.cdn.nextdayvideo.com/veyepar/lca/lca_2015/img/2015_01_12_20_49_02-002.png

chop it into talks:

http://veyepar.ndv.cdn.nextdayvideo.com/veyepar/lca/lca_2015/img/2015_01_12_20_49_02-002-0b.png

OCR, find "9571" and link the image to the page where the data needs to
travel to:

veyepar.nextdayvideo.com/main/C/lca/S/lca_2015/E/9571/

The "chop" step has issues.  There are 2 page layouts (p1 has more stuff in
the header), I have not been consistent about scan techniques (I typically
ask someone else to scan and give me the files) Paper size is not
consistent, printer quality is not consistent.

I don't mind tweaking the pdf to help this process, like add a boarder or
registration marks, and I have been thinking of adding little bar codes if
I can figure out how to do that with reportlab

Here is the code that chops and OCRs and such:
https://github.com/CarlFK/veyepar/blob/master/dj/scripts/addimg.py#L219-L229




On Mon, Dec 14, 2015 at 1:17 PM, Jeffrey Ratcliffe <
jeffrey.ratcliffe at gmail.com> wrote:

> On 14 December 2015 at 14:05, m. allan noah <kitno455 at gmail.com> wrote:
> > Why not just use scanimage -b ?
>
> Or gscan2pdf?
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/sane-devel/attachments/20151214/a5ca6701/attachment-0001.html>


More information about the sane-devel mailing list