宽方位地震数据相干噪声压制技术研究及应用

宽方位地震数据具有以下两个特点: ①远偏移距地震道集中相邻地震道之间的距离明显小于理论道间距; ②横向炮检距越大, 检波线的空间采样越不规则, 相邻道之间的距离变化率越大。这两个特点会导致面波、鸣震等线性相干噪声的时距曲线随横向炮检距的增大而呈双曲线变化, 从而导致常规面波压制技术难以获得理想的噪声压制效果。针对这一问题, 提出了基于十字交叉排列锥形滤波的均值加权相干噪声压制技术, 该方法在十字交叉排列域进行, 首先采用锥形滤波方法提取得到噪声模型后, 对噪声模型进行均值加权处理, 实现信噪分离, 提取噪声并加以压制, 保护有效信号, 得到最终的去噪结果, 提高资料的信噪比。将该技术应用于野外采集的宽方位地震数据相干噪声的压制处理, 结果表明, 在有效去除了干扰噪声的同时, 有效信号基本没有损失, 提高了地震资料的信噪比, 验证了本文方法的适用性和有效性。

For wide-azimuth seismic data, trace interval in gathers at far offsets is much smaller than the theoretical value; the spatial sampling rate tends to be more heterogeneous and trace interval varies quickly at large crossline offsets.Consequently, the time-distance curves of such linear coherent noises as surface waves and reverberation vary hyperbolically with crossline offset; thus, it is difficult to suppress such noises using routine methods.To address this issue, we develop a mean value-weighted approach based on cone filtering to reduce coherent noises in cross spread.The noise model is obtained through cone filtering, followed by mean value-weighted processing, for the separation of signals and noises to improve signal-to-noise ratio.Our approach is validated to be feasible and credible by field data processing as coherent noises in wide-azimuth seismic data could be effectively reduced without impairing signals and thus the signal-to-noise ratio is improved.