Fiting a surface of best fit to a set of data points

[Woland]

Hello all,

I am trying to fit a surface to a 1920x1080 grid of evenly spaced data points. The values are supposed to be more or less uniform (its an image of a uniform white light source). So I would like to fit a plane to it (but maybe a paraboloid if it is not quiete uniform). What method should I use? A least squares approach? I can't seem to find much on google for this. Does anyone have an approach, or perhaps could recommend some literature which covers this?

Thanks!

 

[SW VandeCarr]

If you're just interested in the results rather than method (least squares or MLE), here's some software that might be useful:

http://www.mathworks.com/support/solutions/en/data/1-1AVW5/index.html?solution=1-1AVW5

 

[Woland]

I was able to figure it out . The usual least squares method applies to this. I fit z = Ax + By + C by creating a matrix of sums etc etc. It worked out nicely for the 3D case, same as for the 2D.

Thank you,