应用统计模型的地震信号端点检测方法  

Endpoint Detection of Seismic Signal Using the Statistic of Noise

作　　者：范万春[1] 施仁[1] 孙煜[2] 邱宏茂[2] 李真富[2] 

机构地区：[1]西安交通大学电子与信息工程学院,西安710049 [2]西北核技术研究所

出　　处：《西安交通大学学报》2001年第4期365-369,共5页Journal of Xi'an Jiaotong University

基　　金：国防科研预先研究项目! (5 6 .1.7)

摘　　要：分析了随机信号的相关性 ,建立了噪声的相关性统计模型 .提出了信号自相关系数相似距离的概念 ,证明并讨论了噪声和信号之间的可辨别“自相关相似距离” ,据此总结出基于相关系数的地震信号端点检测方法 ,并给出了算法的具体实现步骤 .阐述了检测阈值的确定方法 ,并对用此方法以及人工检测法得到的实验检测结果进行了比较 .结果表明 ,即使在较低的信噪比条件下 ,采用本算法仍能较准确地检测出地震信号的端点位置 .Based on the analysis of the auto-correlation coefficient, the notion of the distance between auto-correlation coefficients was quoted, and the identifiable distance between the noise and the signal was proved and discussed. And based on the auto-correlated similarity, the auto-detection method of seismic signal endpoints was summed up. The steps of implementation and determining of the threshold were presented in detail. The experimental results that were compared with the method based on artificial detecting show that, the method based on the auto-correlated similarity has higher sensitivity even in a low SNR circumstance.

关 键 词：相关系数 地震信号 统计模型 端点检测 自相关相似距离 端点位置 

分 类 号：P315.61[天文地球—地震学]