基于信号包络线相似性的特定场地地下爆炸识别方法研究

Discrimination algorithm for underground explosions in a fixed site based on the similarity of wave envelopes

摘要: 基于信号包络线相似性，本文提出了特定场地地下爆炸的识别方法，并将其用于一个特定场地地震事件分析，以评估该方法对特定场地爆炸事件的识别效果. 结果表明，在误警率约为6.3%的情况下，事件间距小于15 km的爆炸事件漏检率约为0.5%，事件间距在15—30 km范围内的爆炸事件漏检率约为5.0%. 与波形互相关方法相比，本文方法大幅扩大了有效识别特定场地爆炸事件的空间范围.

Abstract: This paper presented a discriminant algorithm based on the similarity of wave envelopes to discriminate seismic events in the same site.The events in a fixed site were analyzed to evaluate the discriminating performance for underground explosions.The result shows good performance.When the false alarm rate is about 6.3%， the misclassification rate for explosions with inter-distance smaller than 15 km is only 0.5%， and that for explosions with inter-distance of 15—30 km is about 5.0%. Comparison with the result from the algorithm based on waveform cross-correlation indicates that the method in this paper can discriminate explosions with much larger inter-distance.