HOWTO: Scan multiple pages to a PDF file and OCR using tesseract on archlinux

Dependencies:

• simple-scan
• tesseract-svn
• imagemagick
• leptonica
• pdftk
• exactimage
• my bash script

Step 1: Use simple-scan and scan everything to a multipage PDF file.
Step 2: Apply my script to the .pdf file.
DONE

#!/bin/bash

LANG=”deu”
FILE=$1
TEMPS=`echo $FILE | md5sum | cut -c 1-32 `
mkdir /tmp/${TEMPS}
cd /tmp/${TEMPS}

echo “split pages”
pdftk $FILE burst dont_ask output ${TEMPS}_%02d.pdf

for f in ${TEMPS}_*.pdf
do
f=`basename $f .pdf`
echo “Converting ${f}.pdf to tif”
convert -quiet -density 600 -size 210 × 297 -depth 8 “${f}.pdf” “${f}.tif”
convert -quiet -density 150 -size 210 × 297 -depth 8 “${f}.pdf” “${f}.jpg”
done

for f in ${TEMPS}_*.tif
do
f=`basename $f .tif`
echo “Using tesseract on ${f}.tif …”
tesseract “${f}.tif” $f -l $LANG hocr
done

for f in ${TEMPS}_*.pdf
do
echo “adding text to pdf”
f=`basename $f .pdf`
hocr2pdf -n -o ${f}_n.pdf -i ${f}.tif < ${f}.html
pdftk ${f}.pdf background ${f}_n.pdf output ${f}_added.pdf
done

echo "Joining Multipagepdf"
pdftk ${TEMPS}_*_added.pdf cat output "${TEMPS}-ocr.pdf"

dirname=`dirname $FILE`
filename=`basename $FILE .pdf`

echo "writing new pdf file"
mv ${TEMPS}-ocr.pdf ${dirname}/${filename}_ocr.pdf

echo "deleting temp files"
rm -rf /tmp/${TEMPS}

PS: This should work on Ubuntu too