奇异值分解(SVD)实现地震波场分离与去噪新思路

依据不同性质的地震信号（反射波、折射波、直达波、面波、VSP上下行波、多次波、随机干扰等）之间在运动学、视速度和相干性上的差异,借助某种数学变换、SVD分解与重构联合、时域与频域结合的方式,通过这些间接的处理手段,把要提取的目标信号或要剔除的干扰信号转换到一种相干性更好的空间域中,再进行SVD分解与重构,最充分利用SVD滤波技术特点,实现地震波场分离与去噪,而不是直接对信号进行SVD分解与重构来实现地震波场分离与去噪,这样做可有效地避免以往对SVD波场分离与去噪技术应用空间狭窄、有效信号损失严重等缺陷性.

New idea for seismic wave filed separation and denoising by singular value decomposition (SVD)

Based on the differences of propagation characteristics, apparent velocity and coherence in seismic signals for different waves (reflected wave, refracted wave, direct wave, surface wave, up and doen waves of VSP, multiply wave, random signal and so on) , we can transform the singals, which we want to distill or reject to, to one kind of spatial domain with better coherence by some mathematical technique with combined time-frequency domains. Then carrying on the SVD decomposition and restructuring again, we can achieve the seismic wave field separation and denoising. But it is distinguishing obviously with carrying on SVD directly to the seismic wave field separation and denoising. This may avoid the flaws that the technology of wave field separation and denoising via SVD is applicated narrowly and the desired signal loses seriously and so on.