基于保真振幅单程波方程的叠前AVP成像方法

基于共聚焦点技术的叠前AVP（振幅随射线参数变化）分析与常规叠后反演方法相比优势明显，但传统通过褶积和互相关运算来实现的方法依赖于聚焦算子，而在复杂构造区走时计算困难且子波难以精确提取，从而导致了聚焦算子不准确，而且褶积和互相关运算会影响信噪比和分辨率，基于此，本文提出了基于保真振幅单程波延拓算法的叠前AVP成像方法.该方法利用保真振幅傅里叶有限差分延拓算法实现两步聚焦，分别生成共聚焦点道集和网格点道集，既充分利用了保真振幅延拓算法在振幅保持方面的优势，也可以发挥傅里叶有限差分方法对复杂构造区横向变速适应性强的优势，而且两步聚焦过程都不需要聚焦算子，从而解决了传统方法中走时计算和子波提取的问题.模型试算结果表明了方法的正确性和可行性，而针对实际地震资料的试处理结果与传统方法相比具有更高的信噪比和分辨率，表明了方法的有效性.该方法为复杂构造区油气检测提供了一种新的地球物理依据.

Prestack AVP imaging method based on true amplitude one-way wave equation

Prestack AVP (amplitude-versus-ray parameter) analysis based on common focusing point technology has advantages over post-stack inversion method. However, conventional method which is based on convolution and cross-correlation depends on focusing operator. And it is hard for us to calculate travel time and extract accurate wavelet in complex areas, so it would lead to wrong focusing operator. In addition, convolution and cross-correlation will affect S/N ratio and resolution. So in this paper we propose prestack AVP imaging method based on true amplitude one-way wave equation. In this method two focusing steps are realized by true amplitude Fourier finite-difference extrapolation method. And common focusing point gather and grid point gather are separately generated. This method could use not only the preserved-amplitude advantage of true amplitude extrapolation method but also the good adaptability advantage of Fourier finite-difference method to lateral variable velocity in complex areas. And the two focusing steps don′t need focusing operator, so the problems of travel time calculation and wavelet extraction of conventional method could be resolved. Test results of model indicate the validity and feasibility of this method. And we also carry out tests aiming at real seismic data. The processing results of our method have higher S/N ratio and resolution compared with conventional method. This shows the effectiveness of our method. This method provides a new geophysical reference criterion for hydrocarbon detection in complex areas.