Seismic multi-attribute fusion based on pulse coupled neural networks
Li Quanzhong1, Peng Zhenming1, Zhou Jingjing2, Zhang Ping1
1. School of Optoelectronic Information, University of Electronic Science and Technology of China, Chengdu, Sichuan 610054, China;
2. Sichuan Petroleum Geophysical Prospecting Company, CNPC, Chengdu, Sichuan 610213, China
Abstract:This paper presents a method for seismic multi-attributes fusion based on pulse coupled neural networks (PCNN) for solving the multi-solution dilemma in the single-attribute prediction of oil and gas reservoir. First by facilitating PCNN model, we determine the fusion coefficients for all the neurons with the powerful non-linear processing function of neutral network. Then, we obtain the fusion data output of the responding neurons. Finally, the seismic multi-attributes fusion is realized. The salient features of this proposed method contain its succinct structure; its fast calculation and its high quality fusion results. Applications of the clipped multi-attributes data in the northeast Sichuan demonstrate the efficiency and rationality of the method.