filtering and smoothing 2d data

Hi all,

I have latxlon data of sea level pressure but due to its high spatial resolution (50 km), they are very noisy. My task is to smooth (or to filter) these data. I need also to remove noise with defined wavenumbers (bandpass filter)

What is the appropriate way to do this task?

Firstly, To smooth data before filtering, with the use of, for instance, median filter or something else. Secondly, To filter the noise with defined wavenumbers (bandpass filter)?

or to apply only filtering without smoothing?

I would be happy if you provide some code how to smooth or filter with defined wavenumbers.

Thank you very much!

Answer 1

[Image Analyst]

You can use sgolayfilt() in the Signal Processing Toolbox, like in my attached demo.

OR you can use "Alternating sequential filters" like presented in this link: http://perso.esiee.fr/~coupriem/Pdf/cb04.pdf

Answer 2

[Image Analyst]

If you want to smooth an image instead of a boundary of specific latitude by longitude coordinates, then you can use Fourier filtering. I attach some examples that you can adapt.