Script for converting images to text (ocr) for the Noisebridge Archivists group. Requires tesseract. Ubuntu packages: tesseract-ocr and tesseract-ocr-eng

img2txt.sh

#!/bin/bash
#Converts images to text using tesseract (package tesseract-ocr & tesseract-ocr-eng)

function usage
{
	echo "img2txt -i <input directory> -o <output directory> --concat"
}

function concat
{
o=$1
	for f in $o/*
	do
		cat $f >> $o/concatenated.txt
	done
}

function convert
{
i=$1
o=$2
mkdir $o
echo "Converting and placing files into $o"
	for f in $i/*
	do
		test -f $f || continue
		echo "processing file $f... $o/${f##*/}.txt"
		tesseract $f $o/${f##*/} &> /dev/null
	done
	echo "done."
}
inputdir=
outputdir=
concatenate=
while [ "$1" != "" ]; do
	case $1 in
		-i | --input )	shift
				inputdir=$1
				;;
		-o | --output )	shift
				outputdir=$1
				;;
		-c | --concat )	concat=1
				;;
		-h | --help )	usage
				exit
				;;
		* )		usage
				exit 1
	esac
	shift
done

convert $inputdir $outputdir
if [ $concat == 1 ]; then
	concat $outputdir
fi