[R] gaussian model of 3d surface

Martin Maechler maechler at stat.math.ethz.ch
Sat Jul 24 16:41:03 CEST 2004

>>>>> "Alan" == Alan Swanson <swanson at yellowstoneresearch.org>
>>>>>     on Fri, 23 Jul 2004 10:35:08 -0600 writes:

    Alan> Hi all, I'm working with radar remote sensing data and
    Alan> want to measure the height of 3d spikes created by
    Alan> backscatter from bright objects.  My thought is to fit
    Alan> a gaussian density function to each spike and extract
    Alan> the x, y and z location of the peak.  Can anybody give
    Alan> me a hint of how to do this?

If I understand correctly,
you want to fit what is commonly called a "(finite) Normal
Mixture Model" to your 3d data 
and then extract the (estimated) (3d-)means of each component ?

If yes, then consider using package 'mclust'.

Martin Maechler

More information about the R-help mailing list