Wednesday, October 23, 2013

Inspecting cylinders

In the ideal world I’d use a linescan camera to inspect a beer can or jar of pickles. That way there’d be no distortion in my high-res image. The problem with this approach though is that it means rotating said can or jar. It’s a problem easily solved with money, but that’s always in short supply.

The approach many integrators take is to use a number of matrix cameras to image the cylindrical surface. Usually, as in this system developed by UK integrator Adbro Controls, that number is four.

To my way of thinking, that means a lot of distortion out on the periphery of the image. And that problem becomes worse when a short focal length lens is used, as is the case here.

Had I developed the system you’d have seen at least five cameras viewing the OD, each with a long working distance and long focal length lens. Of course, that would make the inspection system much bigger, unless you could come up with a clever way of folding the lightpath, but I have my ways.

In addition to reducing distortion, this would have solved the problem of each light shining directly into the camera opposite. Now in the story, (which I lifted from the most excellent Vision Systems Design magazine,) we’re told each camera is “equipped with a polarizing filter to ensure that light emanating from white light LEDs will not saturate the imager in the camera directly opposite it.”

Say what?

Now if the light is polarized there might be some benefit in using a polarizing filter, but if all they’re trying to do is reduce intensity there are other ways of achieving that result. Like a neutral density filter.

Or am I missing something? Please let me know.


Anonymous said...

Use direction of polarization to filter out unwanted light?

No idea how well it works, but in theory..

Anonymous said...

I'd assume they put filters after the LEDs in one orientation and rotated by 90 degrees in front of the camera. Then direct light is completely filtered, but reflected light is not.