#!/bin/bash

TFILE_BASE="/tmp/ocr.$$.tmp"

#get the last filename and increment the number, keeping the leading zeros
FILENR=`ls -p /home/XXXX/scans | grep -v / |tail -n 1 | cut -b -5`
let "FILENR=10#$FILENR+1"
FILENR=`printf "%05d" $FILENR`
OUTBASE="/home/XXXX/scans/$FILENR"

echo "$OUTBASE"


# Set the language to german if no argument was given
if [ "$#" -ne 1 ]
then
	OCRLANG="deu"
else
	OCRLANG=$1
fi

#echo $OCRLANG

# procedure to enhance the image and finallo do the OCR with a PDF output
do_ocr() {
	# reserve file name for parallel scanning
	touch "$OUTBASE.pdf"

	# create a quality png that is later wapped in the pdf file
	convert $TFILE_BASE.pgm -units PixelsPerInch -density 600 -colorspace gray -colors 8 $TFILE_BASE.png
	# to the actual OCR and store the result in a pdf file
	tesseract -l $OCRLANG $TFILE_BASE.png $OUTBASE pdf quiet 1>/dev/null 2>&1
	# clean up all tmp files
	rm $TFILE_BASE*
}

#echo $TFILE_BASE.pgm

scanimage --mode gray --resolution 600 --contrast 50 > $TFILE_BASE.pgm

do_ocr