Spike Removal

The Spike Removal preprocessor enables you to remove anomalous spiked data from Raman spectra. It achieves this in a two part method. First, it finds spectra with a large difference between neighboring data points (the cutoff parameter). These are processed further: points above the z-score threshold (spikes) are interpolated.

1. Cutoff: only spectra with difference between neighboring datapoints higher than cutoff are going to be processed further.
2. Threshold: the z-score threshold above which points are marked as spikes.
3. Distance: the number of nearby data points (right and left) for interpolation of spikes.

References

• Python code from https://towardsdatascience.com/removing-spikes-from-raman-spectra-8a9fdda0ac22
• Whitaker, Darren A., and Kevin Hayes. “A simple algorithm for despiking Raman spectra.” Chemometrics and Intelligent Laboratory Systems 179 (2018): 82–84.