fitting a double gaussian given one centroid image

[Thex0913]

Hello,

I am currently working on trying to best fit my data points to a double gaussian. Orginally it was fitted to a single gaussian but my professor said it was not good enough. My problem is that I don't think its possible to fit our data points to a double gaussian if we only have one set of data points. This is mainly because we only have one centroid. Is it possible to plot two centroids on one image given that there are two distinct sources of light? (Also forgot to mention that we are fitting the data to the intensity of a laser beam horizontaly across the x coordinate of the centroid and vertically across the y coordinate of the centroid.)

[altenbach]

You don't really say what you mean by "double gaussian".

Seems you have two beams on top of each other. If they differ in width or position, the profile in each direction would be the sum of two gaussian.

This could be fit with a model that includes the xy positions, width, and amplitude of each component.

Why do you think this is not possible?

A single beam could also be distorted such that the horizontal width is different to the vertical width, giving you a 2D gaussian (e.g. if the detector place is not perpendicular rto the beam).

Message Edited by altenbach on 11-30-2009 03:03 PM