基于曲波域模型优化的多次波压制方法在浅地层剖面的应用

Application of multiple suppression based on curvelet-domain modeloptimization in subbottom profiling

浅地层剖面主要用来探查海底的浅部构造和浅部资源,但因地震数据采集时受水深和海底地质条件等的影响,多次波极其发育,并且多次波能量强、频带宽,常规的自由表面多次波压制方法很难取得较好的压制效果。基于此,将地震数据自身褶积产生的多次波模型分为低频段模型和高频段模型,同时,将原始数据分为低频段数据和高频段数据,利用多个自适应相减法优化低频段模型;而对于高频段数据和高频段模型,将其转换到曲波域,比较其不同尺度、不同角度的差异,然后,对高频段模型进行优化,最后,将优化后高频段模型和低频段模型相加,得到最终优化后的多次波模型,再将其从原始数据中直接减去,以达到压制浅地层剖面中多次波的目的。实际资料的应用结果表明,该技术对多阶的海底相关多次波和强界面产生的多次波的压制效果较好,剖面的信噪比得到明显提高,有效信号得以凸显,有利于后期资料的应用。

Subbottom profiling is mainly used to explore shallow structures and shallow resources below the seabed.Owing to the influence of water depth and geological conditions during acquisition,subbottom profiles are usually contaminated with strong multiples in a wide band,which are difficult to remove using conventional surface-related multiple suppression methods.We divide the multiple model generated by the convolution of seismic data into a low-frequency model and a high-frequency model,and divide the original data into low-frequency and high-frequency components.The low-frequency model is optimized by using multiple adaptive subtractions.The high-frequency data and high-frequency model are transformed into the curvelet domain to investigate the discrepancies between them at different scales and different angles and thus to optimize the high-frequency model.The optimized multiple model obtained is subtracted from the original data to suppress multiples in subbottom profiles.Practical application shows good results of suppressing multi-order seabed-related multiples and those multiples generated by the interfaces with large reflectivities.Improved signal-to-noise ratio and enhanced effective signals are advantageous to subsequent data application.