基于相干滤波的相干体图像增强

地震资料的相干分析是一项十分重要的地震解释技术。为了压制沿地层形成的低相干带和随机噪声带来的干扰，相干分析技术需要利用多个地震道和较大的分析时窗来得到相干体。但多个地震道和大分析时窗的采用，会导致所获得的相干体数据的空间和时间分辨率降低。文章提出了一种新的基于相干滤波的相干体图像增强技术，通过对相干体进行相干噪声滤波处理来压制不需要的低相干带和随机噪声带来的干扰，从而使得在进行相干分析时可以使用较少的地震道和较短的分析时窗来获取高分辨率和高信噪比的断层信息。此项技术利用了沿地层形成的低相干带和由断层产生的低相干带方向上的差别，通过局部方向扫描和多项式滤波来滤除不需要的低相干带和随机噪声带来的干扰。实际资料处理结果表明,所提出的算法可以获得真实的、高信噪比和高分辨率的相干体图像，从而提供更丰富的各种地质信息。

IMAGE ENHANCEMENT OF COHERENCE CUBE BASED ON COHERENT FILTERING

Coherence analysis technique plays a key role in seismic interpretation. To suppress lower-coherence-strip (LCS) along horizon and random noise, multi-traces and a large time window are needed in coherence analysis to generate a coherence cube. However the coherence cube will lose its resolution in space-time domain when multi-traces and a large time window are applied. An innovative image enhancement technique of coherence cube based on coherent filtering is presented in this paper. It can suppress undesired LCS and random noise through coherent noise filtering of the coherence cube, so as to obtain fault information with high resolution and signal-noise-ratio by using less seismic traces and a smaller time window in coherence analysis algorithms. Along horizon LCS and differences in directions of LCSs resulted from faults are used to suppress the undesired LCS and random noise through local dip scanning and polynomial filtering. Application of our algorithm to real dataset shows that it can generate real coherent cube image with high resolution and signal-to-noise ratio, thus providing more abundant geological information.