Automatic noise-removal/signal-removal based on general cross-validation thresholding in synchrosqueezed domain and its application on earthquake data
Recorded seismic signals are often corrupted by noise. Wehave developed an automatic noise-attenuation method forsingle-channel seismic data, based upon high-resolution timefrequencyanalysis. Synchrosqueezing is a time-frequency reassignmentmethod aimed at sharpening a time-frequencypicture. Noise can be distinguished from the signal andattenuated more easily in this reassigned domain. The thresholdlevel is estimated using a general cross-validation approachthat does not rely on any prior knowledge about thenoise level. The efficiency of the thresholding has been improvedby adding a preprocessing step based on kurtosismeasurement and a postprocessing step based on adaptivehard thresholding. The proposed algorithm can either attenuatethe noise (either white or colored) and keep the signalor remove the signal and keep the noise. Hence, it can be usedin either normal denoising applications or preprocessing inambient noise studies. We tested the performance of the proposedmethod on synthetic, microseismic, and earthquakeseismograms.
Mousavi, S., & Langston, C. (2017). Automatic noise-removal/signal-removal based on general cross-validation thresholding in synchrosqueezed domain and its application on earthquake data. Geophysics, 82 (4), V211-V227. https://doi.org/10.1190/GEO2016-0433.1