I don't know much about Rician noise in particular, but I have heard is the most difficult noise to reduce because it is not additive and not really Gaussian. Although the researchers seem to use the assumption that it IS Gaussian, and from that approach they can reduce noise. I also know that it is frequently showing up in magnetic resonance imaging.
Check a paper on "The Rician distribution of Noisy Data". Maybe also check Rician noise generator to add the noise.
The noise itself is seemingly linked with magnetism, so I assume that it is the result of a strong magnetic force field, maybe?
See also "Noise estimation from Rician noise corrupted images". The noise reduction properties also seems to depend on the S/N ratio.
As for Rayleigh I don't know and I am interested in anyone else's answer.
