Hi,
I have two 3D models in OpenGL that correspond to each other . I would like to register the model 1 with model 2 and I was wondering if anyone has an idea how to do it. I know how to perform the matching with corresponding point (point-point) but I don’t know how to do it with surfaces.
Thank you very much!
P.S. Please do not refer me to Besl and McKay, it’s very instructive but difficult to implement.
this isn’t an opengl question; this is a computer vision / modelling question.
If you have a model of something then you should be able to work out how that corresponds with another model, but unless you explain how you’ve stored your model, then it’s going to be anyone’s guess as to what to suggest.
Are you seriously saying you have two polygonal models you want to register, or you have two IMAGES of two models (as in a stereo vision system) of the same model??
cheers
John
What is the class of output you are looking for? Is it a rigid, affine or non-linear transformation?
Are you looking for an automatic method or semi-automatic method?
Coco Nut
The model that I have are polygonal model representing the same structure. I’m looking for a rigid method (without deformation / scaling) for the model that will be
transformed.
An automatic method would be great!
What about computing the covariance matrix of all the vertices of each object, and then compute the transformation from one convariance matrix into another matrix?
Coco Nut