Prev: Centre y to x==0 in plot
Next: triangular factorization of sparse matrix
From: Tim Maguire on 15 Apr 2010 10:16
We have a Matlab routine that calculates certain distances based on an acquire image.

We have found though that the images have some curvature to them because of the camera that we use.

Is there a way to remove the curvature from our calculations so that we end up with true Cartesian measurements?

Thanks.
-Tim
From: ImageAnalyst on 15 Apr 2010 10:40
Tim Maguire:
Sure, you can improve it. Take a snapshot of a known thing, like a
grid - crossing stripes with known distances between them. Then
locate where the control points (i.e. crossing points) actually are,
and then specify where the points *should* be. Then use maketform()
to come up with a warping transform. Then use tformfwd() to apply
that transform to "fix up" a bad image.
 | 
Pages: 1
Prev: Centre y to x==0 in plot
Next: triangular factorization of sparse matrix