基于MPI的时频分布的改进及应用

The amelioration and application of time-frequency distributions on the basis of MPI

Cohen类时频分布在一定程度上克服了Wigner Ville分布的不足,特别是交叉项.但在平滑交叉项的同时却降低了时频聚集性,从而本文讨论了Wigner Ville分布的修正方法,进而可提取出不同的地震时频属性.为了提高计算效率,采用了基于MPI的并行计算及时频属性优化方法,取得了良好的效果.

The Cohen's class time-frequency distributions can overcome the shortage of Wigner-Ville distribution, especially the cross-terms. However, the time-frequency concentration is reduced simultaneously. Consequently, its modified method is discussed in the paper, and so it picks up different seismic time-frequency attributes. In order to improve the computing efficiency, we adopt parallel computing on the basis of MPI and the optimized method of the time-frequency attribute, achieving good effect.