Research On Earthquake Signal De-Noising Based On Independent Component Analysis

Earthquake detecting has been an important way to detect underground geologic structure.And earthquake signal is the main information source for earthquake detecting. It is quite a critical step to remove noise from earthquake signal in order to accrue information from earthquake wave.Both earthquake signal and random noise are non-Gaussian and relative independent.This paper mainly uses the ICA method to remove random noise from earthquake signals.Fixed-point algorithm based on negentropy has been studied and applied in earthquake signal de-noising.After serious studying the earthquake features,three main sub-wave modes were built.By simulating the earthquake signal and implementing the simulation in MATLAB,the affectivity was validated.