Date: Nov 13, 2012 5:55 PM
Author: Cameron
Subject: Re: anybody got a complex 2-D median filter script they can share?

"Matt J" wrote in message <ikra1j$ani$1@fred.mathworks.com>...
> "Cameron " <cameron.matthews@yahoo.com> wrote in message <ikr4os$515$1@fred.mathworks.com>...
> >
> > since phase wraps around, I don't think a conventional median filter will necessarily provide any meaningful data.

> ============
>
> Well, you could apply the filter in regions where the signals are supposed to be continuous (i.e., for phase, not near pi and -pi) and then fill in the holes with some sort of extrapolation.
>

> > SO - I guess an addendum to this request for help would be:
> >
> > if I median filter the amplitude,then separately median filter the phase or angle and then resolve the two datasets to a complex number, is the math in fact non-valid?

> ===============
>
> What math are you referring to? What equations need to be satisfied? The main concern I would have is if your complex data is going to be passed through some sort IFFT operation. In that case, errors in the median filtering could propagate spatially throughout the entire result.


Hey you still there? took me long enough... anyhow, I think "math" was a poor choice of words on my part. I guess if a median filter is analagous to an edge preserving low pass filter, then I am concerned that there is no value to the edges in a low pass of a phase signal. what I really want to preserve are the phase components that aren't noise and that aren't impacted by the different grazing angles. its sort of like two cameras looking at different angles of the same picture, and the cameras get phase data that lets the light be sampled coherently. whats of value in that coherence between the two views? I don't know if a median filter will get it.