基于匹配追踪算法的立井爆破信号时频特征提取

在对立井爆破过程中井壁振动信号进行准确采集的基础上，采用匹配追踪（matching pursuit，MP）算法建立Gabor原子库，实现了爆破信号重构分量和残差分量的有效分离。通过信号短时傅里叶变换（short time Fourier transform，STFT）和魏格纳分布（Wigner-Ville distribution，WVD）对重构信号的时频分析及霍夫(Hough)变换时频聚集性对比，验证了WVD时频分析的优势。结果表明:MP-WVD组合信号分析方法对信号变化的适应性强，能够精确捕捉信号的局部细节，适合用于爆破信号非线性时频特征的提取过程。可为爆破方案调整和参数优化提供参考。

Time-Frequency Feature Extraction of Shaft Blasting Signal Based on Matching Pursuit Algorithm

Based on the accurate acquisition of the vibration signal from shaft lining during the process of vertical shaft blasting, the Gabor atomic database was established by using the matching pursuit (MP) algorithm to realize the effective separation of the reconstruction component and the residual component of the blasting signal. The advantages of WVD time-frequency analysis were verified through the comparison of time-frequency analysis and Hough transform time-frequency aggregation of the reconstructed signal by STFT and WVD. The results show that MP-WVD combined signal analysis method having features of strong adaptability to signal change and accurate capture on the local details of the signal is suitable for the extraction process of nonlinear time-frequency characteristics of blasting signal. It can provide reference for blasting scheme adjustment and parameter optimization.