Seismic Signal Processing by Wavelet Based Multifractal Analysis

摘要

Fractal geometry and wavelets are a new and promising approach to analyze and characterize non-stationary signals such as seismic signals, ECG, stock prices, etc. Short-term Fourier transform analysis imposes the assumption that signals are stationary over small temporal segments. Such an assumption is inappropriate for seismic signals because conditions inside mines undergo constant change. Because they do not impose this assumption, wavelet-based techniques are proving useful for processing seismic data. In this paper, we propose a wavelet-based method to analyze seismic signals, which is superior to many conventional methods used in the industry today. The fractal dimension and other measures are calculated by using wavelets, in which time is irrelevant. Another measure, Multifractal Spectrum is computed with the help of a scaling exponent. Using this strategy, our method distinguishes between a seismic signal and a noisy signal or truly chaotic signal by fractal dimension. Further our result shows that the larger fractal dimension indicates a rock burst in mines. The approach introduced here should be useful in the analysis of other non-stationary biological signals.