Searching \ for '[EE]: Project idea??? Impossible???' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=project+idea+impossible
Search entire site for: 'Project idea??? Impossible???'.

Exact match. Not showing close matches.
PICList Thread
'[EE]: Project idea??? Impossible???'
2001\03\26@234744 by Richard Sloan

flavicon
face
I have been considering a device for a company.

They have around 10-20 different "certificates" that need to be sorted, I would like to do some type of recognition of them and sort them automatically. The device needs to be cheap and I feel doing something like a CCD and digitizing them would be too costly for this project.

I envision a real dirt cheap proximity sensor type arrangement where the document would pass over it and the sensor would just sense light and dark regions similar to bar codes. The actual image it may see however is more text and graphics.

I would also like the device to be able to handle the document passing the sensor at a nonlinear rate, or rather possibly at a unknown and variable rate, as if someone pushed the document past it by hand.....

Keep in mind cheap! Sounds almost impossible, but I know someone out there has some ability to shed some light on the issues.

Thanks!

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email spam_OUTlistservTakeThisOuTspammitvma.mit.edu with SET PICList DIGEST in the body


2001\03\27@004825 by Dale Botkin

flavicon
face
On Mon, 26 Mar 2001, Richard Sloan wrote:

{Quote hidden}

There was a thread not long ago regarding vending machine bill acceptors
doing something very similar.  Assuming the certificates have no existing
bar codes, I think you may be a lot better off building a motor-driven
feed (DC motor running at constant speed, rubber wheels, opto sensor for
detecting the leading edge) and trying to develop some sort of "signature"
pattern for each document.

I'd look at a defunct cassette tape transport or VCR or something for the
parts for that...  you can't beat free parts and a little plywood or sheet
metal.

Dale
---
The most exciting phrase to hear in science, the one that heralds new
discoveries, is not "Eureka!" (I found it!) but "That's funny ..."
               -- Isaac Asimov

--
http://www.piclist.com hint: To leave the PICList
.....piclist-unsubscribe-requestKILLspamspam@spam@mitvma.mit.edu


2001\03\27@005402 by Brandon Fosdick

flavicon
face
Richard Sloan wrote:
>
> I have been considering a device for a company.
>
> They have around 10-20 different "certificates" that need to be sorted, I would like to do some type of recognition of them and sort them automatically. The device needs to be cheap and I feel doing something like a CCD and digitizing them would be too costly for this project.
>
> I envision a real dirt cheap proximity sensor type arrangement where the document would pass over it and the sensor would just sense light and dark regions similar to bar codes. The actual image it may see however is more text and graphics.
>
> I would also like the device to be able to handle the document passing the sensor at a nonlinear rate, or rather possibly at a unknown and variable rate, as if someone pushed the document past it by hand.....
>
> Keep in mind cheap! Sounds almost impossible, but I know someone out there has some ability to shed some light on the issues.

Sounds a lot like a hand scanner. They use a ball or rollers like a mouse to
track the motion of the scanner over the page. In your case instead of moving
the scanner, you're moving the page.

Have they thought about just scanning the documents and doing image/ocr
recognition?

-Brandon

--
http://www.piclist.com hint: To leave the PICList
piclist-unsubscribe-requestspamKILLspammitvma.mit.edu


2001\03\27@022201 by D Lloyd

flavicon
face
part 1 1914 bytes content-type:text/plain; charset=us-ascii
Hi,

I'm sure it can be done; I worked on a (partially successful - prototype
worked great, demand not so good) project involving the distinguishing of
disc rotations in electromechanical meters (without breaking the seal). We
shone IR onto the disc edge and monitor the reflected power to gain a
signature. It even worked for discs without the customary "black mark."

Good luck with your system,
Dan






(Embedded     Richard Sloan <.....rsloanKILLspamspam.....THEMINDFACTORY.COM>KILLspamspam.....MITVMA.MIT.EDU>> image moved   27/03/2001 05:51
to file:
pic17549.pcx)





Please respond to pic microcontroller discussion list
     <
EraseMEPICLISTspam_OUTspamTakeThisOuTMITVMA.MIT.EDU>
Sent by:  pic microcontroller discussion list <PICLISTspamspam_OUTMITVMA.MIT.EDU>


To:   @spam@PICLISTKILLspamspamMITVMA.MIT.EDU
cc:
Subject:  [EE]: Project idea??? Impossible???

Security Level:?         Internal


I have been considering a device for a company.

They have around 10-20 different "certificates" that need to be sorted, I
would like to do some type of recognition of them and sort them
automatically. The device needs to be cheap and I feel doing something like
a CCD and digitizing them would be too costly for this project.

I envision a real dirt cheap proximity sensor type arrangement where the
document would pass over it and the sensor would just sense light and dark
regions similar to bar codes. The actual image it may see however is more
text and graphics.

I would also like the device to be able to handle the document passing the
sensor at a nonlinear rate, or rather possibly at a unknown and variable
rate, as if someone pushed the document past it by hand.....

Keep in mind cheap! Sounds almost impossible, but I know someone out there
has some ability to shed some light on the issues.

Thanks!

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email KILLspamlistservKILLspamspammitvma.mit.edu with SET PICList DIGEST in the body






part 2 165 bytes content-type:application/octet-stream; (decode)

part 3 105 bytes
--
http://www.piclist.com hint: To leave the PICList
RemoveMEpiclist-unsubscribe-requestTakeThisOuTspammitvma.mit.edu


2001\03\27@054733 by Mg

flavicon
face
You know the length of the certificates so you can use one light/detector
sensor to work out how fast it is sent through. Then you can use another
light/detector to take some voltage samples from the reflected light,
average these and compare the average to a stored table in a PIC which
contains the already worked out averages of the 20 different types of
certificates. I'm only a novice at electronics at the moment so could this
work???
-Mg

--
http://www.piclist.com hint: To leave the PICList
spamBeGonepiclist-unsubscribe-requestspamBeGonespammitvma.mit.edu


2001\03\27@082624 by spam

flavicon
face
Hi Richard

I have seen a nice little holographic optical component that uses
the laser-doppler principle to measure the rotation speed of a table
tennis ball. HO components are cheap to make and this would
probably solve your problem.

Kent

{Quote hidden}

--
http://www.piclist.com hint: To leave the PICList
RemoveMEpiclist-unsubscribe-requestspam_OUTspamKILLspammitvma.mit.edu


2001\03\27@084035 by Quentin

flavicon
face
Pretty much the kind of thing I do designs for. (Printing, labels and
paper handling, etc.) This is a basic principle behind document or card
(like Lotto) readers.
The documents, is that standard already printed? Or can some marks be
added to the documents being printed?
If you can, then print something like this on the side of the document:

XSXSXSXSXSXSXSX
MSMSMSMSMSMSMSM

M=Mark
S=Space
X= Mark or Space

Use two sensors. Sensor1 reads the Mark/Space, sensor2  the X/Space.
When sensor1 picks up a Mark, sensor2 reads whether the X is a mark or
space.
Shift that into a register and compare with knows values and voila!

Look at Industrial fibre optic sensors for a sensor for your needs. Try
Omron or Keyence (.com for web pages).

If it is already printed documents, you will have to look at something
like OCR or Machine vision unfortunately ($$$$$$$$$$$$!!!!!).

Quentin

--
http://www.piclist.com hint: To leave the PICList
RemoveMEpiclist-unsubscribe-requestTakeThisOuTspamspammitvma.mit.edu


2001\03\27@111809 by rottosen

flavicon
face
I like this idea.

Use several sensors each looking at a diverse set of points on the
certificate. Maybe 8 of them each on an A/D input to a PIC. Develop a
signature for what each certificate looks like. No scanning is required
but the sensors and the certificate will need to be in a jig of some
sorts to get repeatable alignment for consistent results.


-- Rich


D Lloyd wrote:
{Quote hidden}

--
http://www.piclist.com hint: To leave the PICList
EraseMEpiclist-unsubscribe-requestspamEraseMEmitvma.mit.edu


More... (looser matching)
- Last day of these posts
- In 2001 , 2002 only
- Today
- New search...