A Seismic Fault Recognition Method Based on Region Energy Algorithm
Lei Chen,Yanqing Liang,Guanglei Qi,Ting Zhang
DOI: https://doi.org/10.1142/s0218001424500150
IF: 1.261
2024-08-25
International Journal of Pattern Recognition and Artificial Intelligence
Abstract:International Journal of Pattern Recognition and Artificial Intelligence, Volume 38, Issue 10, August 2024. Fault recognition is a difficult problem in seismic exploration data interpretation, and there is still no solution both well in terms of accuracy and signal-to-noise ratio. To solve this problem, based on the region energy algorithm, a novel fault recognition method is proposed, which determines the direction of fault tracking based on region energy when identifying fault points. First, the third-generation coherence cube algorithm is adopted to calculate the coherence attribute of the seismic data volume. Then, fault tracking is performed on each seismic section. When conducting fault tracking, the seismic sample is scanned and identified one by one. If it is a fault point, it is assigned to the corresponding fault in the connected area, and then, track along a certain direction of the current pixel point in the front left, directly ahead, or front right direction. The selection of the tracking directions is based on the energy of the corresponding area in the direction. The direction with the highest energy is tracked in the direction until the complete fault is tracked or the stopping condition is reached. If the point is not judged as a fault point, a certain distance is tracked down continue and the path is stored temporarily. If a fault point is tracked, the tracking path is classified as a fault, otherwise return to continue scanning. When all the sample points on the seismic section are scanned, the fault tracking on the corresponding section is completed. Subsequently, the fault points are fitted using the least squares fitting algorithm, and the fault line is obtained. Finally, comparative experiments were conducted on actual seismic data, and the effectiveness of the novel method was validated.
computer science, artificial intelligence