17 Dec
2009
17 Dec
'09
8:17 a.m.
On Thu, 2009-12-17 17:06:36 +0100, Jan-Benedict Glaw <jbglaw at lug-owl.de> wrote:
On Mon, 2009-12-14 14:15:30 +0000, Philip Pemberton
<classiccmp at philpem.me.uk> wrote:
[...]
Hmpf, the script was stripped. Here it is, inline:
#!/usr/bin/env sh
NETPBM=pbm
TESSLANG=deu
HOCR2PDF=/home/jbglaw/paper/hocr2/HocrConverter.py
# 2 = pixel/inch, 600x600 DPI
TIFF_296=2
TIFF_282=600
TIFF_283=600
OUTDIR="__out__.$$"
rm -rf "${OUTDIR}"
mkdir "${OUTDIR}"
# Split
mkdir "${OUTDIR}/split"
for i in *.tiff; do
TIFFNAME="`basename "${i}" | sed -e 's/\.tiff$//'`"
tiffsplit "${i}" "${OUTDIR}/split/${TIFFNAME}-"
done
# Convert to bw/grey/colour NetPBM format for unpaper
mkdir "${OUTDIR}/singlepage_pbm"
for i in "${OUTDIR}"/split/*.tif; do
TIFFNAME="`basename "${i}" | sed -e 's/\.tif$//'`"
convert "${i}" "${OUTDIR}/singlepage_pbm/${TIFFNAME}.${NETPBM}"
done
# Beautify scans
mkdir "${OUTDIR}/singlepage_unpaper"
for i in "${OUTDIR}"/singlepage_pbm/*."${NETPBM}"; do
PNMNAME="`basename "${i}" | sed -e
's/\.'"${NETPBM}"'$//'`"
unpaper --pre-rotate 90 --input-pages 1 --output-pages 2 --layout double "${i}"
"${OUTDIR}/singlepage_unpaper/${PNMNAME}-%d.${NETPBM}"
done
# OCR all the stuff
mkdir "${OUTDIR}/hocr"
for i in "${OUTDIR}"/singlepage_unpaper/*."${NETPBM}"; do
PNMNAME="`basename "${i}" | sed -e
's/\.'"${NETPBM}"'$//'`"
tesslanguage="${TESSLANG}"
export tesslanguage
ocroscript rec-tess --tesslanguage="${TESSLANG}" "${i}" >
"${OUTDIR}/hocr/${PNMNAME}.hocr"
done
# Generate TIFFs from beautified NetBPM images
mkdir "${OUTDIR}/unpaper_tiffified"
for i in "${OUTDIR}"/singlepage_unpaper/*."${NETPBM}"; do
PNMNAME="`basename "${i}" | sed -e
's/\.'${NETPBM}'$//'`"
convert "${i}" "${OUTDIR}/unpaper_tiffified/${PNMNAME}.tiff"
tiffset -s 296 "${TIFF_296}"
"${OUTDIR}/unpaper_tiffified/${PNMNAME}.tiff"
tiffset -s 282 "${TIFF_282}"
"${OUTDIR}/unpaper_tiffified/${PNMNAME}.tiff"
tiffset -s 283 "${TIFF_283}"
"${OUTDIR}/unpaper_tiffified/${PNMNAME}.tiff"
done
read -p "Fix images now and press enter" FOO
# Generate single page PDF files
mkdir "${OUTDIR}/single_pdf"
for i in "${OUTDIR}/unpaper_tiffified"/*.tiff; do
TIFFNAME="`basename "${i}" | sed -e 's/\.tiff$//'`"
"${HOCR2PDF}" "${OUTDIR}/hocr/${TIFFNAME}.hocr" "${i}"
"${OUTDIR}/single_pdf/${TIFFNAME}.pdf"
done
# Generate final PDF
pdftk "${OUTDIR}/single_pdf"/*.pdf cat output "${OUTDIR}/book.pdf"
MfG, JBG
--
Jan-Benedict Glaw jbglaw at lug-owl.de +49-172-7608481
Signature of: Tr?ume nicht von Dein Leben: Lebe Deinen Traum!
the second :
Also, does anyone know of an app that can take
the PDF file, OCR it
and then insert the text as a background layer while leaving the
image alone? I'm pretty sure Acrobat can do this, but like most
Adobe software, the price tag is somewhat... eye-watering. "If you
have to ask how much it costs, you can't afford it."
The script may be used under the terms of the GNU General Public
License Version 3.