中文摘要：

即使采用分辨率很高的双曲Radon变换,对速度各向异性发育介质及长偏移距情况下的地震数据,其Radon域内能量仍不收敛.为了克服此难题,我们在Radon变换的积分路径中考虑了非双曲走时的影响,通过引入非双曲时差公式中的各向异性非椭圆率η参数,可以准确描述出长偏移距条件下来自同一层位的时距曲线,并推导了由偏移距、慢度、非椭圆率三参数控制的积分曲线正反变换公式,我们称之为各向异性Radon变换.离散化求解时,各向异性Radon变换是时变的,频率域快速算法已不适用,本文采用了最优相似系数加权Gauss-Seidel迭代算法,保持其计算精度的同时也有较高的计算效率.将此方法应用在模型数据以及实际长偏移距海上地震数据的多次波压制处理中。收到了较好的处理效果.

英文摘要：

For the seismic data acquired in the anisotropic medium and with large offset,although we use the high-resolution hyperbolic Radon transform,the energy of the Radon domain is still not convergent.In order to overcome this problem,the influence of the anisotropic factor is considered in the integral path of Radon transform.By introducing the anisotropic anellipticity parameters η from non-hyperbolic moveout formula,the travel-time curve can be accurately described in the horizontal layer for large offset.We derive the integral equation of forward and inverse Radon transform with three parameters,which are the offset,the slowness and the anellipticity.For time-varying of anisotropic Radon transform,the fast algorithm in the frequency domain is not applicable in the process of discretization.This paper adopt the optimal semblance weighted Gauss-Seidel iterative algorithm which not only keeps the high resolution but also has the high calculation efficiency.In the examples of the anisotropic model and real marine seismic data with large offset,this method shows the effective and efficient application to demultiple.