This one must have been fun to come up with because it’s got it all. There’s hardware, firmware, networking, and server scripts all working together to create a filing, scanning document center for your business. The best part is that [Janis Jakaitis] was tasked to do this as part of his job (we’re sure there’s a bunch of IT guys shaking their heads at this statement, but it sounds like fun to us!).
The goal was to use an existing document scanner to create PDFs which are then stored in a filing system on the network. Of course it needed to be automatic. The first big issue was that the scanner was USB only, and when connected to a USB-to-LAN bridge the buttons on the device no longer functioned. [Janis] put together an Arduino circuit that added that button, as well as a display to show the status of your scan job.
The next issue is getting the filing system to recognize the document as a unique file. The solution here is to generate a unique barcode label that can be affixed to the page before scanning. Since this is a standalone setup, it was tricky to get the label printer to spit out a unique label. He already had the Arduino working with the scanner, so [Janis] decided to use it to drive this barcode job as well. It calls to a Lua script running on the server, which then pushes the next unique code to the printer.
Tie it all together and you get the demo video after the break.
[youtube=http://www.youtube.com/watch?v=tdxwOx3m-fM&w=470]
Is the label printer that really necessary? Unless they want to reconcile the hard copy with the scanned version?
Label bear the information where scanner will save the pdf file. Later accountants use the same label on the paper document to link document on the server with document (transaction) in accounting system (ERP).
As this process covers incoming invoices from many different suppliers with no any document standardization between them – OCR and automatic document data entry in ERP won’t work.
From the vid, the label gets scanned on the first page of the document, telling the server where to save it. (And, in turn, by scanning the invoice with a phone/etc to see where it was saved)
And from clicking around on the website, it seems that it is indeed for reconciling invoices to an SAP server.
See flow chart at http://3.bp.blogspot.com/-TXSNDKvy6VQ/TkOpbTkQEyI/AAAAAAAAAOA/7XT-iYbWFEE/s1600/Slide1.JPG
Actually it doesn’t matter what ERP is used. Simply when accountants process paper copy (enters information in ERP) they click on field “path to file” and scan bar-code on the paper with hand-held scanner. That’s how path to already scanned PDF is entered in ERP. It is open to any system if you can spare document form field for file path.
Is something like this available commercially? I really need it.
Commercial solutions are from 40k eur. This one costs 4 …
Hmmm…. weekend project:
Sacrifice old A4 printer and scanner to do it all in one pass, Arduino based of course :D
This is not a weekend project. I had to design it with off the shelf parts and make it easy to replicate. We have two locations using this solution already and I hope to roll out some more next year. So Rube Goldberg contraptions (despite that I really like them) was not and option. Even connecting the Arduino to the scanner button and LCD will void the scanner warranty and was not acceptable.
You didn’t understood me.
I said, this would give a nice weekend project to ME.
Using and old scanner and some printer to make it all in one pass only.
I run a scanning service. I can think of several past customers that this would have been their best solution instead of what they where trying to implement with all the data entry they did duplicated by the scanning company. This looks like a pretty good integrated solution for processing the invoices.
Contact me, I will be glad to help with advice if you decide to build you own :)
A place I worked at about 3 years ago was doing this using a product that is now called Kofax Capture which did the scanning and OCR that our product then picked up to allow for checking of the OCR data and approval and finally export to the financial system saving on data entry costs without requiring the paper document once it was scanned.
We actually use Kofax Capture and Hummingbird DM to capture and store our information respectively. They cost a fortune but scale so nicely. Being able to run multiple scanners capturing, cleaning up, OCR’ing, and indexing the paperwork with Kofax is a dream, then having it export to Hummingbird so that hundreds of people can work and use the document simultaneously is another god-send. Professional solutions tend to scale wonderfully. Our setup is designed to import around 100k pages _per day_. Right now we do around 5-7k a day and we’re barely making it work.
It always confuses me that people get upset at how expensive this sort of solution is. Yeah, we spend probably close to $75k to implement, and yeah, there’s a profit line in there, but its not like you’re just burning the money. It’ll pay for itself in just reduced management costs alone. Not to mention support, compatibility, and compliance.
I don’t doubt this solution works well for its intended audience, but it wouldn’t work for our situation. Always good to see some alternatives.
Built systems like this for customers, municipalities and law firms mainly. Using existing Xerox hardware and M$ SharePoint/linux servers. The tagging option is _really_ god if you have large amounts of stored printed material. If you run a law firm it’s a must almost.
It’s an efficient system to use especially if you got multiple locations.
Thank you! It was really fun and challenge to build. Especially printing data matrix code on the printer which have not this capability built in. Another good part was to go through several iterations making project simpler and simpler until all we have are two buttons and handheld scanner :)
It seems from this feedback that there is already a lot of interest in a similar affordable product.
Market opportunity anyone?
I hardly imagine product consisting for 90% from other products – HP, Zebra, Honeywell. This is not a product, this is solution. It can be bought from the company I’m working on :)
I write software for Windows platforms and you can say that my end product is probably 99% Dell(hardware)and MS (Operating System). This does not mean I can’t charge for my modifications (my software). If you have the design and aren’t bound by corporate copyright law (work-for-hire), you could, with a little effort make a reasonable second income. $40 in parts could easily generate $500 in profit and still be welcomed by the customer when compared to the $$$ alternatives. I used to work for a document imaging company and the expenditure of $500 on a project would be considered negligable.
Great solution. I liked how you solved this problem. :) Does anybody know what i the name and the author of the book he is referring too? He mentioned “And then the engineer said…” but I cannot find that book in Google. I have a 5 year old son and would like to read it to him.
Unfortunately it’s in Russian http://www.kodges.ru/23527-i-tut-pojavilsja-izobretatel.html
HP has a higher end scanjet that attaches to the network.
For a higher price :)