Automatic event detection in low SNR microseismic signals based on multi-scale permutation entropy and a support vector machine
| Autor: | Xin-Ming Lu, Hong-Mei Sun, Yongquan Liang, Rui-Sheng Jia, Yan-Jun Peng |
|---|---|
| Rok vydání: | 2016 |
| Předmět: |
Feature vector
02 engineering and technology 010502 geochemistry & geophysics computer.software_genre 01 natural sciences Signal Physics::Geophysics Geochemistry and Petrology Least squares support vector machine 0202 electrical engineering electronic engineering information engineering 0105 earth and related environmental sciences Event (probability theory) Mathematics Microseism Noise (signal processing) business.industry Pattern recognition Scale factor Support vector machine Geophysics 020201 artificial intelligence & image processing Artificial intelligence Data mining business computer Seismology |
| Zdroj: | Journal of Seismology. 21:735-748 |
| ISSN: | 1573-157X 1383-4649 |
| DOI: | 10.1007/s10950-016-9632-2 |
| Popis: | Microseismic monitoring is an effective means for providing early warning of rock or coal dynamical disasters, and its first step is microseismic event detection, although low SNR microseismic signals often cannot effectively be detected by routine methods. To solve this problem, this paper presents permutation entropy and a support vector machine to detect low SNR microseismic events. First, an extraction method of signal features based on multi-scale permutation entropy is proposed by studying the influence of the scale factor on the signal permutation entropy. Second, the detection model of low SNR microseismic events based on the least squares support vector machine is built by performing a multi-scale permutation entropy calculation for the collected vibration signals, constructing a feature vector set of signals. Finally, a comparative analysis of the microseismic events and noise signals in the experiment proves that the different characteristics of the two can be fully expressed by using multi-scale permutation entropy. The detection model of microseismic events combined with the support vector machine, which has the features of high classification accuracy and fast real-time algorithms, can meet the requirements of online, real-time extractions of microseismic events. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full text from SpringerLink