‎08-27-2007 03:03 AM
‎08-28-2007 07:27 AM
‎08-28-2007 09:43 AM
Thanx for the prompt reply!
I've seen this. Did not look very deep into it, 'cause I thought that it was not what I need.
I need a perspective transformation in a dynamic environment. The perspective coordinates are calculated dyncamically given some kind of position, angle and distance. And I need to transform an image from perpendicular transformation (rectangle) to perspective transformation (parallellogram)
It seems to me, that the procedure you mentioned, is some kind of learning to determine the perspective distortion.
Or am i mistaken?
‎08-29-2007 03:20 AM
Hi,
Yes that is correct. The calibration works based on a predefined grid and then this grid is read in with the camera. Based on this grid the learn VI will calculate the compensation needed and store this with the Grid Image (you can save this image to disk with the calibration information attached).
Next the calibration info from the grid is copied to the incoming image and used trough the analysis for the images etc. This more static and more ment to get correct measurement result for shape size and distances etc.
NI-Vision does not contain any functions to apply the calculation to the image data directly
Regards
Karsten