基于动态图像变形的PP与PS波层位直接匹配

多分量地震资料的矢量偏移、多波地震资料的联合解释与反演均需要估算纵横波的速度比,实现纵波与转换横波在时间或深度域的匹配.基于DTW,本文实现了一种适用于PP与PS波直接匹配的动态图像变形算法.该算法分为三个部分:首先,使用二阶对称动态规划算法逐样点递归计算PP与PS波走时或深度的误差累积和;其次,在以误差累积和为目标函数的回溯阶段设定变形窗,并在纵横波速比约束的变形窗内递归回溯搜索匹配路径;最后,根据最大相关系数判定准则在匹配路径中确定最佳匹配路径,获得使PP与PS波匹配的拉伸或压缩时移量.利用所获得的拉伸压缩时移量计算纵横波速度比就可以实现PP与PS波之间的匹配.模型与实际陆上多分量地震资料测试结果表明:该方法具有较高的匹配精度,且对于信噪比、相似度较低的多分量地震资料,该方法也能产生较好的匹配效果.

PP- and PS-waves matching directly based on dynamic image warping

P- to S-wave velocity ratio estimation as well as PP- and PS-waves matching in time or depth domain are the basic procedures for the imaging, inversion, and interpretation of multi-component seismic data. Based on Dynamic Time Warping (DTW), we propose a three-step dynamic image warping algorithm which can be used to match PS- to PP-waves directly in time or depth domain of post-stack sections. With the second order symmetrical dynamic programming method, we first compute the total alignment errors of the double-way travel time or one-way depth for PP- and PS-waves. The warping window on total alignment errors is calculated with the maximum and minimum of the P- to S-wave velocity ratio to constrain the searching zone. We then search the optimal path within the warping window by using the recursive backtracking method. A maximum correlation coefficient criterion is applied to determine the best warping path. We finally calculate time- or depth-shifts between PP- and PS- images and match PS to PP post-stack section with the best shifts or the P- to S-wave velocity ratio calculated from the shifts. Our method is tested by using synthetic and field data from Xin Chang oil field. The results show that our method is effective and accurate in matching the PP- and PS-events, and could be further applied to the seismic data with low signal-to-noise ratio or weak similarity in different components.