Seismic signal detection and first arrival pickup based on mutual information

In this paper the fundamental theory of information science-Information Theory is introduced to the analysis and processing of seismic signals. Based on mutual information, a novel seismic signals detection and first arrival pickup method is proposed with the principle and detail algorithm. The computation results on real seismic data show that mutual information in the sliding observation window can indicate the existence of the seismic signals to be detected. Through the pickup of the rising-break or the peak point of mutual information, the arrival time can also be effectively estimated. Especially, the presented method can be used in the processing of the non-classical seismic signals such as continuous stochastic signals and random impulse sequences. When analyzing these kinds of signals, the signal noise ratio is usually very low and seismic waves will change apparently with the transmit distance. So it is difficult to analyze these signals by classical detection methods, but the proposed method can provide some useful results in these situations.