中文摘要：

为改善强随机噪声背景中地震勘探资料信噪比（S/N）,提出了混沌振子算法;该算法可实现对未知同相轴的确定.本文的时空域双曲滤波（hyperbolic ti me-distance relation filter,HTDF）是上述算法的后续处理,即处理得到存在的弱同相轴.HTDF的理论基础是最小平方滤波和混沌振子算法.为说明滤波过程与效果,对于包括一个反射同相轴的地震记录进行了多道滤波因子计算.结果表明,与一个输出道相应的双曲滤波因子组具有复杂的时频域结构,突出的特点是振幅谱“波动式陷频”表现;弥补或消除这个不足的方法是与输出道相应的多个滤波因子之间存在的振幅谱“互补”性.利用相同随机噪声背景下的共炮点理论记录,经与域双曲速度滤波、褶积滤波处理方法比较,在振幅谱、子波初至、零相位性、S/N改善等方面,HTDF的滤波效果优于另外两种.最后探讨了该滤波技术的机理,包括同相轴内各子波间的相似性,以及对组内因子实施的多次有效相加处理.

英文摘要：

The chaotic vibrator method （CVM） were applied to improving the signal-to-noise ratio （S/N） of seismic prospecting data with strong random noise, and ascertain the unknown event- fix. In this paper, the hyperbolic time distance relation filtering （HTDF） is presented, which is a next signal processing followed CVM for the enhancement of weak events. The HTDF is based on the both least square filtering algorithm and chaotic vibrator theory. In order to describe the HTDF＇s process and result, multi trace filtering-factors are calculated for the seismic record contained a reflecting event. The simulation experiment results show that the HTDF＇s factor of a out-trace has very complex construction in time frequency domain, and a prominent characteristics is the undulate fall-frequency phenomenon （UFFP）. The complementation characteristics of the amplitude spectrum of many filtering-factors relative an out-trace is a way by which decreasing or eliminating UFFP. The simulating experiments indicate that the effect of the HTDF is better then hyperbolic velocity filtering in domain and convolution filtering. Finally, the mechanism of the HTDF technique is discussed, which includes comparability of the wavelets in an event and the effective multi-addition processing for the factors in a group.