Ricean noise

[MatthewWhiting]

Enno Middelberg would like to use Duchamp to examine cubes with Faraday depth as the third axis. The problem with doing this is that the noise in such a cube is not Gaussian, but Ricean. This means that the way the noise parameters and the probability of each pixel being noise or source needs to be done differently.

Investigate the Rice distribution and how one estimates parameters, and incorporate into the cube stats functions.