[Help-gsl] Beginner trying to use nonlinear least squares

[James Fitzsimons]

Hi all,

I need to compute a nonlinear least squares solution for my computer
vision project. I am trying to work out the ego motion
(x,y,z,roll,pitch,yaw) of my camera system given two observed point sets
that represent the same features at time t0 and t1.

I think I need to be using the functions for multidimensional nonlinear
least-squares fitting. The thing I am having trouble understanding is
how to provide the function(s) to be minimised.

If I am trying to compute the partial derivative of x for the i'th point
and my function looks something like this:

delta_Ci = (u0 + (f * (Xi - delta_x) / Zi)) - (u0 + (f * Xi / Zi))

(where u0 and f are constants).

How would I construct a gsl_multifit_function_fdf for this?

Any help you can provide would very much appreciated!

Regards,
James Fitzsimons

signature.asc
Description: This is a digitally signed message part