This is a partial answer to some of your questions
On Tue, Feb 01, 2005 at 05:00:34AM +0000, Stephen Tait wrote:
> Yargh, this has to be the worst job ever. A client has landed a rush job on
> us in that they need about 6GB of TIFF files (from scanned documents)
> converted into OCR'd PDF's within the next 3-4 days. Yay.
>
> The problem is the images we've been given are utter cack. Firstly, because
> of Kodak's seeming inability to produce decent drivers, the pages are
> inexplicably surrounded by huge black borders. Because the papers are all
> of different sizes and aspect ratios, there is no way I can do an
> ImageMagick/IrfanView batch crop on them. I found the GIMP's autocrop tool
> worked perfectly on the sample images I downloaded to my home machine, so
> I've set about making a batch process script for it.
>
I managed to learn to use perl-fu reasonably quickly. I'm not such a big
fan of the "Lots of Isolated Silly Parentheses" like stuff found in
scheme.
It might be easier to do the general file management stuff in perl than
in scrpt-fu as well.
> Secondly, as well as mixed borders, we also have mixed TIFF formats. Most
> of the images are in standard 2-colour Group4 fax compression, but they're
> interleaved with 256 colour LZW TIFF's and 24bit colour JPEG TIFF's.
>
> So I'm frantically trying to come up with a script-fu for GIMP that can
> take all of this into account. I'm currently attempting to use the glob
> plugin to take care of the fact we have around 200 directories full of
> images (so it'll process /path/to/images/*/*.tif), although I'm not sure if
> the glob in GIMP 2.2.0 supports this (although I'll have access to 2.2.2 at
> work).
>
> Secondly, what is a drawable? I've seen vague references to it in the
> script-fu guides I've skimmed over, but I am at a loss to see where they
> fit into the bigger picture. Do I need to convert a file into a drawable,
> run the autocrop on that, and then pipe that into an image somehow?
>
I think it's something you can draw to - like a layer or a mask or a
channel, etc
> Thirdly, does anyone know if there's some way to specify a "save this TIFF
> in the same format/colour depth you opened it in" option? If not, I can
> save all output as full colour LZW's and let Adobe Capture do the
> resampling, but it's going to cost a fortune in discspace and I doubt the
> client will be happy either.
>
One issue you will have, is that gimp doesn't have a true bilevel mode
(a *big* shortfall if you ask me). It will load the faxg4 images as
greyscale and the best it can do is to save as a paletted image wth a
palette containing only two colours. You'll then have to convert back to
tiffg4 using ImageMagick - but this might be possibly to do within your
perl script or maybe scheme aswell - I've not looked far enough.
As far as telling the difference, maybe there is a gimp scripting
function to count the number of unique colours?
> Sorry to sound like a complete I-don't-know-what-I'm-doing, but I probably
> don't I'm chowing my way through HOWTO's in between trying to write the
> script as we speak, but would greatly appreciate any advice the community
> could give me, if only to be able to get to bed before the sun comes
> up Yay for still working at 5am
>
> P.S. if anyone has any other suggestions for applications that'll batch
> remove arbitary black borders (be they Linux or Windows, although I don't
> have the power to buy stuff myself) they too would be welcome.
>
>
--
David Purton
dcpurton@chariot.net.au
For the eyes of the LORD range throughout the earth to
strengthen those whose hearts are fully committed to him.
2 Chronicles 16:9a
Attachment:
signature.asc
Description: Digital signature