Don
Do you have the raw data?
The pixels are highly rectangular, so it would be important to have the raw data without the repeated pixels in the x direction. (sure, I could probly add some code to extract it, but why?)
Each image column is already quite smooth, so what you need to do is reduce the x dimensions to a set of unique slices and then interpolate to generate the new, smooth, rows at the present scale to end up with to square pixels again.
Here's a quick attempt using 2D circular comvolution with a Gaussian of 0.1% width in y and 3% width in x. I am sure once you can post the raw data, things will fall into place just fine.
Message Edited by altenbach on
07-09-2008 02:55 PM