hardware & Software for image processing

[Vindhyachal.Takniki]

1. We have a application in which a machine is displaying data on a graphical data like date, time, Count, pass/fail & 3 others.

2. There is no way output from machine, so what I am thinking of purchasing a image processing hardware which get pics of that lcd & then process it to get the info from it & store it in PC for further process..

3. What is right hardware & software for this?

[wiebe@CARYA]

Depending on your exact requirements, this will work on any cheap laptop with it's webcam or requires a super PC cluster with state-of-the-art camera.

The key here is what you mean by "then process [the image]". Taking a snapshot and storing it? Getting a colour? Or let a 15 level NN analyse it?

And at what speed? Taking an image, and analyse it during the next 3 hours? Or 30 frames per second?

The choice of camera also depends on the optical system requirements. Webcams are pretty limited in their setup, while vision camera's allow a multitude of lens setups. Any vision expert will tell you the quality of the image is key, so spend enough time thinking about it. Also note that cheap webcams may only provide a compressed stream. This might have compression artefacts that could ruin your image.

If you just want to interpret some values with OCR, let's say once a second, I would think a mid-priced PC with mid-priced webcam will do just fine...

[Bob_Schor]

I agree with Wiebe's analysis, and think that this is a really kludgy (is that a word?) way to do this (but realize you may have no alternative).

I'm assuming the device you are describing is an "all-in-one" package that has internally-connected sensors, proprietary software and algorithms, and is either incredibly sophisticated/expensive or ridiculously cheap so that re-engineering the task using a Laboratory Virtual Instrument Engineering Workbench (think of the previous four words as an acronym) would provide a cost-effective (and better) solution.  Of course, you need someone with some LabVIEW experience and expertise, but I'm guessing that because you are posting here, you have that ...

Bob Schor

[wiebe@CARYA]

I had this situation once, where the DUT was in a ESD testing environment. Nothing could be wired to it, even though the devices had serial or other means of communication. There only was a glass window, and some pore guy had to write the values down (on paper of course) every 10 minutes...

Kludgy seems to be a word...