We have a client who wants to use a USB camera to take pix of customers, and also a USB scanner to scan ID's and Checks.
It may be possible, with the right camera (i.e., good resolution and autofocus) to use the camera for ID's and Checks instead of the scanner.
There is no OCR component to either of these, we simply want to kick off a child process, capture the pic and store it - something along the lines of
makefoto /vars/clients/pic/041003300233.jpg
Of course, if the functionality could be made available for linking into our own c code, that would be a real plus...
Either way, "makefoto" would place the camera (suitably installed, of course) into streaming mode, display the picture on the screen and wait for any keystroke; upon seeing the keystroke, it would freeze the current picture, and ask "OK?" with possible answers of Y - save the picture in the given file, exit N - loop back to streaming pictures wait for another keystroke Q - exit immediately
The scanner interface would be substantially the same, except that it would only take one "picture" at a time, but it would still need to display the pic and see if it was okay or not. Something like
makescan /vars/clients/ids/041003300233.jpg
The specific camera is more or less immaterial, as long as it's got autofocus and produces reasonably good quality pictures. They are currently getting acceptable results (on a Windoze box) with several different cameras including an Intel CS0430 and a WatchPort V (used to be 3Com), but at the cost of a good bit of mousing and clicking, plus having to type the file name... I know the Intel is listed on the spca50x site on sourceforge as having a USB driver, so it might not be too hard to run.
The scanner they like so far is a Scanshell, which is specifically sized for IDs and checks, but I don't know the model number, and haven't received information on the chipset or hardware/firmware engine yet - there may be a SANE driver already done, or it may be not too difficult to do. Our client has talked to the people at Scanshell, and they seemed to be willing to assist with any specs or technical information needed. If we go with this scanner, we won't necessarily need the autofocus on the camera.
One last thing - this needs to be in console mode, not X-windows. They're using zgv to display their existing photos right now in console mode (not in an x terminal), and it uses svgalib, so I assume the console mode is not likely to be a problem.
If you know of such an application, or you can produce it (or them), please post contact information here. Or if you're just interested in discussing it, jump right in!
Thanks, all.