基于第二代Curvelet变换的地震资料随机噪声衰减

Seismic data random noise attenuation based on the second generation Curvelet transform

摘要: 噪声衰减是地震资料处理中的关键问题之一。根据Curvelet变换对含有光滑边界的二维二阶连续可微函数所具有的稀疏表示性能,给出了Curvelet变换域地震资料随机噪声衰减的阈值方法;并给出了基于地震资料中随机噪声是独立同分布的高斯白噪假设条件下的阈值估计方法。通过合成数据和叠后实际数据算例,对该方法的有效性进行验证。结果表明,Curvelet变换不仅可以很好地衰减随机噪声,并且能较好地保持有效信号。

Abstract: Noise attenuation of seismic data processing is one of the key questions that can not be ignored.According to the nonlinear approximation property of the Curvelet transform in objects with piecewise C₂ edges,a threshold method of random noise attenuation in seismic data is proposed based on the Curvelet transform.And an estimation approach of the threshold value is presented on the assumption that the random noise is independent and identically distributed Gaussian white noise.The synthetic data and post-stack real data processing results confirm the effectiveness of the proposed method.