几种重排算法在地震信号处理中的实验分析

阐述了时频分析技术中重排算法的基本思想，给出了部分重排算法的局部能量重心表达式；设计了两个数值实验，利用几种重排算法对单道地震信号进行了计算和分析比较。结果表明：重排算法不仅能有效抑制交叉项，同时也能提高时频聚集性，从而获得更理想的局部信号特征。其中以重排Spectrogram分布的时频聚集能力最佳，尤其适合于突出主频信号；重排平滑伪Wigner-Ville分布同样也能够获得较理想的时频效果，但稍弱于前一种方法；重排伪MH分布的时频聚集性相对较差，没有对分散的能量进行有效的重新分配处理。重排算法因其突出的时频聚集能力，应用在地震剖面中可以得到更加细致的瞬时频率特征。

Experiments and analysis in seismic data processing with several kinds of reassignment methods

The main idea of reassignment method in time-frequency analysis is presented and some expressions used to compute the accurate location of the local energy gravity center are given; then two numerical experiments are presented to show how these reassignment methods work individually as used to process single-trace seismic data. Based on the experiments, some conclusions are drawn that the reassignment methods can not only effectively reduce cross-terms, but also lead to a good time-frequency concentration. Compared with others, the reassignment spectrogram has the best potential for ideal time-frequency concentration, especially when the dominant frequencies are required to be underlined. The reassignment smoothing pseudo Wigner-Ville distribution can also achieve perfect effects though it is a little worse than the former one. The pseudo Margenau-Hill distribution is the worst of the three, since it fails to relocate the dispersive energy to the calculated center. A pretty precise instantaneous frequency section can be produced when the reassignment method is applied to deal with seismic signals.