【摘要】 本文研究了利用v-SVC支持向量分类机对天然地震与人工爆破事件进行自动识别的有效性及适用性。首先对波形记录利用不同的小波基函数进行4层小波包变换,然后对变换得到的最后一层小波包系数提取出香农熵特征,然后利用v-SVC对提取得到的香农熵特征进行训练和识别。实验结果证明,利用v-SVC对与人工爆破事件进行识别,可以取得很好的识别效果,识别率最高可达98%。更多还原

【Abstract】 The v-SVC support vector machines for classifying earthquake and explosion is implemented in this paper and experiments are carried out for validity and applicability of this kind classifier.Firstly,the transform of 4-wavelet packet is applied to seismic wave recordings.Secondly,the last layer coefficients of wavelet packet from the transform are employed to extract Shannon entropy.Then,the feature is supplied to a classifier of v-SVC support vector machines for verifying validating the capabilities of the Shannon entropy feature.The results showed that the feature of Shannon entropy is a very good feature to discriminate earthquake and explosion,getting a high recognition rate 98% in our experiments.更多还原