Analysis of Seismic Signal and Detection of Abnormalities

Seismic signals are ground vibrations used to detect seismic events. However, seismic signal captured from sensors is distorted signal contains noise and makes actual event detection difficult. In most cases, external noise such as manmade or any heavy vehicle vibration always overlaps with the seismic reflections over time. The presence of noise in the seismic signal makes it difficult to determine the magnitude at which the seismic events have occurred. The aim of our study is to process the signals received from seismic sensor and identify it as seismic events signal and non-seismic events signal based on the magnitude. The authors propose a robust noise suppression method using bandpass filter, IIR Wiener filter and event detection using recursive Short-Term Average (STA)/Long Term Average (LTA) and Carl Short Term Average (STA)/Long Term Average (LTA). The proposed study determines reference magnitude to distinguish seismic and non-seismic activity. The projected study is based on the analysis of seismic signal received from single sensor and sensor networks (SN) and determines the magnitude to distinguish seismic and nonseismic events and time of an actual earthquake event. The experimental dataset is a broadband seismic signal from BSVK and CUKG station sensors located at Basavakalyan, Karnataka, and the Central University of Karnataka respectively. The proposed approach helps to extract the information about preseismic event, actual seismic event, post-seismic event activities and identify the abnormal pattern that supports to detect heearth’s activities before the actual seismic event.