Is there a document which explains how IMAQ Learn calibration template works ? If not could anyone point me to good source for a theory behind nonlinear mapping/transformation ? I want to implement similar algorithm to camera image calibration, but to a different type of data (not image). Currently I use IMAQ vis to interpolate nonlinear mesh/grid.
Digital Image Processing (Gonzalez)
Is considered one of the best resources in image processing concepts.You shall find all what you seek.
There are many calibration techniques and we can use them according to the application demand. Let me just give you an idea on grid calibration. You get spcial grid targets from vendors for this purpose. But just take a a4 sheet printout with dots(evenly spaced). Take a image (ensure perpendicularity). Take the object that needs to measure without disturbing the camera and take a snap.
Measure the dot spacing along the X-axis and Y-axis and feed the values to co-ordinate system.vi. This is just an over view and very general. If you want details and we can discuss further. And waleed we meet after a long time.
Thanks for asking. Well I left the IMAQ nonlinear calibration to interpolate my data. What I was asking for really was how the nonlinear algorith (used by functions LV_LearnCalibrationTemplate and LV_ConvertRealWorldToPixel in NIVision.dll ) works, as I wanted to write similar one for my need. What is the theory behind it (nonlinear 2D mesh interpolation) ? Anyway. I had more urgent stuff to do, and since the IMAQ solution worked I left it:
Hi all, I'd like to revive this post. I too would like to know how NI designed this particular algorithm. My team would like to interface the LabView image correction algorithm with an outside program, and it would be extremely helpful if we could know exactly how the LV_LearnCalibrationTemplate function operates. Any information at all on the specific algorithm would be greatly appreciated.
Thanks in advance.
As this is a relatively old thread now, I suggest that you will receive a much better response if you post it in a new thread. You can then link this forum post in your new thread.