We have written an EPICS driver for USB Video Class Cameras (UVC) which encompass most consumer webcams. We have tested it with several cameras ranging between 50 - 200 $
BNL NSLS-II Beamline Controls Area Detector repositories - epicsNSLS2-areaDetector
in the ADUVC repository. Also character recognition can probably be done in OpenCV w/ some form of machine learning, and we have been developing a Computer Vision plugin in the ADCompVision repository that could be a
good start.
Has anyone developed an Optical Character Recognition plugin for
AreaDetector? The potential use case is to have a camera aimed at an
instrument front panel, and digitize the display to extract the
instrument's displayed output. I've already anticipated the potential
problems associated with multiplexed displays that may introduce
artifacts into the digitized images, so I don't want to attempt to
develop something that may be a failure anyway.
Are there any classes of inexpensive low resolution cameras supported by
AreaDetector? Like consumer grade webcams, for instance?
Yes, it's a hack, but easier than the alternatives in at least one
known case.