低信噪比地震数据图像结构引导去噪方法与应用

地震噪声影响着复杂断裂结构与储层的刻画, 有效压制噪声并保留地震资料中的结构信息, 是解释性处理的关键。基于保边光滑原理, 在地震数据处理中提出并引入图像结构引导滤波方法, 实现自适应地保留断层、裂缝以及储层反射等关键信息并有效压制散射与随机噪声, 提高地震数据信噪比。正演模拟数据的去噪试验结果表明, 相比常规滤波方法, 图像结构引导滤波方法具有更好的保边光滑特性; 与构造导向滤波方法相比, 该方法不需显式提取层位走向和倾角信息, 具有较高的计算效率。实际三维资料的应用结果表明, 该方法能有效压制不同类型噪声、提高信噪比和有效反射同相轴的连续性; 处理后对应的相干属性可有效增强断裂、储层的分辨能力, 更好地描述断裂系统的展布, 有利于后续地震资料解释工作。

In view of the negative effect of seismic noises on complicated structure and reservoir interpretation, effective noise reduction without impairing seismic structural information is crucial to interpretation-oriented seismic data processing. Based on the principle of edge-preserved smoothing, a self-adapting image texture guided filtering method is developed to suppress scattered and random noises in poststack seismic data and improve the signal-to-noise ratio, and meanwhile seismic reflections related to faults, cracks, and reservoirs could be preserved effectively. Model tests show that this method performs better than conventional filtering methods in edge-preserved smoothing; compared with structure-guided filtering, this method has high computational efficiency because it is unnecessary to extract strike and dip information explicitly. The field application to a 3D seismic survey shows the advantages of this method in preserving structural details of faults and cracks, attenuating various noises to improve signal-to-noise ratio, and improving lateral continuity and seismic resolution. Coherence attributes calculated using seismic data after image-guided denoising show enhanced resolution of fractures and reservoirs, which is beneficial to subsequent seismic interpretation.