基于拉普拉斯算子的叠前逆时噪声压制方法

Pre-stack reverse-time noise suppressing method based on Laplacian operator

低频噪声是叠前逆时偏移成像的主要问题。分析了叠前逆时噪声的产生机制和基于拉普拉斯算子的图像去噪机理,采用拉普拉斯算子分别在叠加成像域、共成像点域和共炮点域分别进行了叠前逆时噪声的压制处理试验,并与高通滤波压噪方法进行对比。数值实例研究表明,偏移速度的不准确是造成叠前低频逆时噪声的主要原因,此外正演子波波形与数值模拟记录波形的不匹配和应用相关逆时成像条件是引入低频成像噪声的次级成因。另外,在不同域的基于拉普拉斯算子的低频噪声压制效果均较为明显,其中在共炮点域的去噪效果最佳,且其去噪效果要优于常规滤波方法,去噪后的剖面上地层细节特征更加清晰,共成像点道集上的水平同相轴能准确地凸显出来。

Low frequency noise is the main problem of the pre-stack reverse-time migration section.The figure noise removal mechanism based on Laplacian operator and the raising mechanism of the pre-stack reverse-time migration are analyzed.Laplacian operators are used to carry out the pre-stack reverse-time noise suppressing experiments in stacking image domain,common imaging domain and common shot domain respectively,and the results are compared with high-pass filter method.The numerical simulation results show that the low-frequency reverse-time migration noise is mainly caused by the inaccuracy of migration velocity,and other reasons of low-frequency reverse-time migration noise are the inconsistency of wavelet form between numerical simulating wavelet and numerical records and the application of the correlation type reverse-time imaging condition.The noise suppressing effects based on the Laplacian operator in the above different domains are obvious,and the result in common shot domain is the best.The above methods in three domains are better than that of the conventional high-pass filter method,the formation detail characteristics of the processed section are clearer,and the horizontal cophasal axes of the common imaging gather are correctly depicted.