If you're not opposed to a hardware solution, there are lots of MICR check scanners available. They typically have a USB or serial interface.

It's been a long time, but I recall that they require proprietary software, though you might be able to automate it with Perl.

To do what I am doing would cost (hardware + service) $2300+. I may still have to go that route. I just would really like to see something like this so people can use it without paying big bucks. *sigh*

One thing to consider, if you are looking to automate a check reconcilliation system, is that not all the checks you write will come back to you. More and more of my accounts are going over to electronic funds transfer. What appears on my bank statement is a reference to the check number I wrote, with no returned piece of paper (for example "DES:CHECK PYMT ID:3347") to close the audit trail. I suspect that I can get the paper check, it will just cost the Earth and half the Ocean to do it.

Sigh. This Modern Age....

MICR readers have grown up and much like Mag Stripe readers are available as 'Keyboard Wedge'. Basically they read the check and send the data as a keyboard device. These are also available for under $250.


grep

One dead unjugged rabbit fish later

Thing is those are one at a time. My co workers have about 1k per cubicle- plus i suspect 25k+ in archive that would be nice to get rid of.

i don't know if the office can afford that kind of expense. For the heavy duty hardware.

I think bank of america will let you view a photo of your check. (I wonder about doing something so someone can wget/rip those off their account online.)

I'm doing this for an accounting offices. This one has thousands upon thousands that are being gone through by. .. hand- sometimes they *have* to- a lot of these have funny glitches that some normal MICR check readers would not do

I've got something a million times better and faster. A lot of these scanners will turn a scan to a pdf first. I use that to scan as much as i want to one document to send somewhere. That way I dont have to name and save each scan.

Then I am using pdftohtml to extract images. And imagemagick to do any transforms etc. The set of cli tools for imagemagick like convert are awesome.

Check out what I do.. (this is bash)

#!/bin/sh
rm -rf ./tmp; # clean previous run

# this assumes that the checks are scanned upright, bottom towards lef
+t margin on pdf, one per page
# this assumes you have scans in ./scans
# in pdf format each may have multiple checks, one per page.

mkdir ./tmp
mkdir ./tmp/scans; # incoming pdf with multiple checks
mkdir ./tmp/pngs; # documents for storage prepped

# copy incoming to staging area
cp ./scans/*pdf ./tmp/scans/;

# rip images and resize , crop etc
find ./tmp/scans/ -iname "*pdf" -exec pdftohtml -q -c -zoom 10 '{}' \;
mv ./tmp/scans/*png ./tmp/pngs/;
rm -rf ./tmp/scans
# format the pngs for storing
mogrify -rotate -90 -chop 0x450 ./tmp/pngs/*png;

# Basically all this does is take the pdfs
# and put images in tmp/pngs that are sized properly and right side up
+.

# STEP 2
# get images from pngs and turn into micr strip files
# ready for gocr

# copy to micr
rm -rf  ./tmp/micr;
mkdir ./tmp/micr;
cp ./tmp/pngs/*png ./tmp/micr/;

mogrify -chop 0x660 ./tmp/micr/*png;

# convert the isolated micr for gocr
find ./tmp/micr/*png -exec convert '{}' '{}'.pbm \;
rename png.pbm pbm ./tmp/micr/*pbm
rm ./tmp/micr/*png
[download]

Then I use perl and gocr to rename the files, sort, archive, whatever.. etc etc...