Signal separationΒΆ
The core idea is to separate Bragg-peaks from amorphous signal. Peaks are always seen as positive outliers on top of background noise or amorphous signal. The calculation of the background has gradually improved over time:
Initially, it was a median filter on the azimuthal direction after a 2D intergation. The obtained background curve was often noisy. This median filter has been replaced by a sigma-clipping to enforce a normal distribution of the background. Finally the sigma-clipping has been implemented using 1D integrators, avoiding pixel splitting and mixing together rings.
The reconstitution of the Wilson plot is also presented.