A New Approach for the Detection and Processing of Seismic Signals

Abstract

Prognosis of earthquakes has always been an important area of interest for scientists worldwide, albeit with a very little success so far. The seismic signal represents the inelastic response of the ground structures. Seismic signals are mainly characterized by a ground vibration of finite amplitude and relatively short duration. This paper studies the possibility of earthquake detection by tracking the seismic signals by a sonar equipment and the locating the epicentre by a satellite system. The analyzed signal is composed by a deterministic constituent (the seismic wave) perturbed by noise (random signals) from various sources. Both linear methods such as Short Time Fourier Transform (STFT), and quadratic methods as Wigner-Ville Distribution (WVD) and ambiguity function are used. These methods can be successfully applied to seismic signal analysis and are principally intended for detection and tracking of undersea epicenter. Improvements are achieved by the use of filtering methods, similar to signal processing in the frequency domain. In contrast with classical filtering methods, weighting function is replaced in time-frequency domain by the Weyl function, which characterizes the system simultaneous both in time domain and in frequency domain. In this paper, it will be shown that joint time-frequency techniques could be successfully used to analyze the seismic signal to improve qualitative and quantitative extraction of the most important information.