Faulted feature detection in a tight sand gas reservoir based on steerable pyramid voting Available to Purchase

A steerable pyramid voting method for multi-scale and multi-direction fault skeleton extraction is proposed in this paper. The process can be applied to conventional attributes. In this method, the attribute is enhanced in the optimal direction and in multi-scales using steerable pyramid decomposition, reconstruction and filtering. The process can detect faults, sub-faults and possible fractured zones. However, due to noise effect and low imaging quality, it is very common to have ‘broken’ features which leads to be fuzzy and noisy. The tensor voting based on steerable pyramid is proposed. That is, the relatively high-value faults is used as the seed points to calculate the voting values. All seed points are processed to obtain the voting attribute which can be a further enhancement for the steerable pyramid filtering. This method is applied to the faulted feature identification in a tight sand gas reservoirs in Sichuan Basin. The results show that this method effectively improves the faulted features from seismic. The faulted zones have a good consistency with geology and wells.