Sunday, August 5, 2012

Lining up the linescan cameras

I’ve done a few web inspection projects that needed an array of linescan cameras, though I suppose, strictly speaking, it’s a 1D array. In every case, alignment was an absolute PITA . (If you don’t know that acronym, Google it.) So you’ll understand why I read and re-read “Custom Calibration Supports Linescan System Design” on the Vision System design website.
If you use the link I provided, you’ll find a detailed discussion of how two companies worked on a method of simplifying linescan camera alignment. There were two major components of this: a calibration target that produced data for tweaking camera positions, and a five degrees of freedom camera mounting system. If you’re working with multiple linescan cameras, I encourage you to read this very carefully.

You might also notice that the article is dated December 1st, 2011. I can’t think how I missed it first time around, but I am very pleased Andy Wilson (the Editor of VSD) has taken to providing encore presentations of the most interested articles published. Without that I might have been forever struggling to line up my linescan cameras.


Anonymous said...

Anyone know good method for laser triangulation system calibration?

Anonymous said...

The best approach is probably the same as with line scan systems: take a 3D calibration/aligment target that covers both cameras, then match the two views in 3D and determine the 3D translation vector. Only you have to align both the camera and the laser.

Dlls for calibration and matching of 3D laser triangulation setups can be found e.g on the following website:

Anonymous said...

Does anyone sell similar camera mounts like in article?