论文部分内容阅读
在地震资料处理中可以用中值滤波器来衰减相干波场,在VSP资料处理中衰减下行波就是其中一个例子。该滤波器沿波场某个“方向”对全道集做滤波,未滤波记录减去滤波后的记录就得到最终结果。这种中值滤波方法可称之为“相消法中值滤波”。自动估算记录中相干波场的慢度,根据各点的慢度将该方法推广,以时、空变方式应用于整个地震记录。 中值滤波器是非线性滤波器,因此要了解其性能比了解线性滤波器性能要困难得多。不过,仍有许多方法可以用来分析中值滤波器性能:(1)对具体的时间序列用伪传递函数;(2)中值滤波器对简单地震模型的响应;(3)分析中值滤波器对模拟诸如叠加剖面上断层之类的终止波的阶跃响应。依据这些简单方法可直观了解这些滤波器的性能,同时也给出了一种半定量测量方法。振幅逐道变化时中值滤波器性能下降与线性滤波器的性能下降相类似。若(在t域)加一个低通滤波器,且在记录上斜着做中值滤波可改善滤波器的性能(就信号畸变而言)。中值滤波器对阶跃的响应受背景噪音强弱的影响。中值滤波器引起的阶跃畸变接近于强噪音背景下相应的线性滤波器的阶跃畸变。
Median filters can be used to attenuate coherent wavefields in seismic data processing, one of which is to attenuate the descending wave in VSP data processing. This filter filters the entire channel set along a “direction” of the wave field, resulting in the final result of the unfiltered record minus the filtered record. This median filtering method can be called “the median filtering cancellation method.” The slowness of the coherent wave field in the record is automatically estimated. According to the slowness of each point, the method is generalized and applied to the entire seismic record in a time and space-dependent manner. The median filter is a nonlinear filter, so understanding its performance is much more difficult than knowing the performance of a linear filter. However, there are still many ways to analyze the performance of the median filter: (1) a pseudo-transfer function for a specific time series; (2) a median filter’s response to a simple seismic model; (3) an analysis of median filtering The step response of a simulator to a terminating wave such as a fault in a superimposed section. Based on these simple methods, we can intuitively understand the performance of these filters, but also gives a semi-quantitative measurement method. The performance degradation of the median filter when the amplitude varies by one channel is similar to that of the linear filter. Adding a low-pass filter (in the t-th field) and doing a median filter diagonally on the record improves filter performance (in terms of signal distortion). The response of the median filter to the step is affected by the strength of the background noise. The step distortion caused by the median filter is close to the step distortion of the corresponding linear filter in the context of strong noise.