地震事件实时检测中野值的辨识与替代

Identification and elimination of the outliers in seismic events real-time detection

对地震事件做出快速而准确的判定是地震预警的关键技术,然而背景噪声中的干扰信号可能导致地震事件的误判。结合STA/LTA方法的特点,提出一种基于短时间窗的野值辨识与替代方法。该方法具有野值辨识准确,替代值能够较好保持信号原有的变化趋势的特点。经仿真分析和地震事件波形检验,该方法能有效消除脉冲干扰,对短时连续干扰也有一定弱化效果,能降低地震事件的误判率。

Fast and accurate judgment for seismic events is the key technology of earthquake warning,while the undesired signal in background noise would cause misidentifications.Combined with the characteristics of STA/LTA method,a method of identification and elimination of outliers based on short-time window is proposed.Using this method,outliers can be identified accurately,and the variation tendency of original signal can be kept better by the replace value.Through simulation analysis and examination using waveforms of seismic events,it shows that this method can effectively eliminate pulse interference,weak the short-term continuous interferencea,and reduce the rate of misidentification about seismic events.