Resolving interfering events through cyclic sampling and median filtering
Analysis of a seismic event in prestack data is of ten impeded by interfering events. In such cases it is of interest to remove the interfering events in such a way that neither the amplitude nor the phase of the target event is severely distorted. This is a crucial consideration in amplitude versus offset (AVO) analysis, for example. A common method of tackling this problem is to apply an f-k filter to remove the interfering events. This requires that the target event and the interfering events show sufficient difference in moveout (or dip) at all offsets, which is not always the case. For example, in a common midpoint (CMP) gather the difference in moveout between interfering primary and multiple events may be too small, especially for short offsets. Sharp edges to the filter in f-k space are likely to introduce artifacts in the data (Gibb's phenomenon).