Drawing optical faults to the picture

[Zealous]

Hi all!

I'm using LabVIEW7.1 and imaq Vision 8.x

I coded a program which draws line between picture coordinate and real world coordinate. The information for this task is obtained from calibrated image. I use "convert pixel to real world"-function to obtain real world coordinate and then program draws line between asked and obtained coordinates. The problem is that output image cannot be right. Optical faults usually affects on border of image and in the center of image aberration is minimal. Now we see from OutPutImage that aberration is low upper left corner and that cannot be right.Corrected image seems to be right thou...

I hope you share advices and suggestions conserning my problem

See attacments below

[BruceAmmons]

It looks like your program is just off by one grid point.  If you look at the center, the indicated error is one dot space in each direction.  If you adjust all your errors by subtracting one grid spacing, I would bet it will look exactly like what you want.

In the general case, measure the error at the center of the image and assume it should be zero.  Subtract this error from all future measurements.

Bruce

Message Edited by Bruce Ammons on 03-31-2008 10:27 AM

[Zealous]

Thank you so much!

I did like you adviced. My program calculates and draw errors follows:

1. First Asks real world coordinates in image center, program receives "Subtractor"

2. Then program use same function in loop using user defined x/y -step and obtain real world coordinates and subtract both x y values with      "Subtractor"

3. Finally draws line between asked coordinate and realworld (Subtracted) coordinate

Now result images look like it should be. Still one question remains, why I have to do this in hard way? Or are we just fooling ourselves? It just feels odd to manipulate NI function answers to recieve wanted results :S

Anyway thanks again your kind reply 🙂

Some results as a attachment

Message Edited by Zealous on 03-31-2008 11:37 AM

[BruceAmmons]

I'm not sure how your program is doing everything.  There might be a simpler route, such as setting the center of the image as the origin of your real world coordinates.  I think image calibration is more of a relative than absolute, so when you need absolute you need to define your own reference.

Bruce

[Steve.Briggs]

Anybody know the best way of going about the same optical aberration effect in the images above from Zealous?