College of Geo-Exploration Science and Technology, Jilin University, Changchun 130026, China

Multiple prediction and subtraction techniques based on wavefield extrapolation are effective for suppressing multiple related to water layers. In the conventional wavefield extrapolation method, the multiples of the seismic data are predicted from the known total wave field by the Green function convoluted with each point of the bottom. However, only the energy near the stationary phase point has an effect on the summation result when the convolutional gathers are added. The research proposed a stationary phase point extraction method based on high-resolution radon transform. In the radon domain, the energy near the stationary phase point is directly added along the convolutional gathers curve, which is a valid solution to the problem of the unstable phase of the events of multiple. The Curvelet matching subtraction technique is used to remove the multiple, which improved the accuracy of the multiple predicted by the wavefield extrapolation and the artifacts appearing around the events of multiple are well eliminated. The validity and feasibility of the proposed method are verified by the theoretical and practical data example.

Supported by the National Science and Technology Major Project (No.2016ZX05026-002-003), and the National Natural Science Foundation of China (No. 41374108)