Effect Analysis of Random-weighted Equipoise Multi-channel Adaptive Matched Filtering on Regular Interference Removal

Abstract

Equipoise multi-channel adaptive matched filtering is an effective method to remove regular interference signals from the original seismic records. In order to weaken the dependence of data on orthogonality restrictions, some channels around the target are usually selected as equipoise channels. In order to accurately extract the wavelet shape, the selected equipoise channels need to be as few as possible. The fewer equipoise channels, however, will lead to unsatisfactory filtering effect because of data non-orthogonality. In view of this defect, we develop conventional equipoise multi-channel adaptive matching via randomly selecting channels away from the target as the equipoise channels and assigning different weight factors to different equipoise channels simultaneously. This method may weaken the data non-orthogonality and reduce the influence related to the number of equipoise channels. In order to verify the effectiveness of this method, numerical experiments and analysis of different models have been carried out. The results show that in the case of a small number of matching channels, the effects of the random or random-weighted equipoise multi-channel adaptive matched filtering are better than the conventional one. The random-weighted has the best effect.