基于分块字典学习理论的地震数据去噪  

Denoising of Seismic Data Based on Block Dictionary Learning Theory

作　　者：周俊捷 吴相伶 李文杰 李静和[1] ZHOU Junjie;WU Xiangling;LI Wenjie;LI Jinghe(College of Earth Sciences,Guilin University of Technology,Guilin 541004,China)

机构地区：[1]桂林理工大学地球科学学院,广西桂林541004

出　　处：《CT理论与应用研究（中英文）》2022年第5期557-566,共10页Computerized Tomography Theory and Applications

基　　金：广西自然科学基金(区域检测数据驱动下探地雷达隧道地质预报智能化解译技术研究与应用(2021GXNSFAA196056);小波与曲波组合域探地雷达数据噪声与背景杂波压制技术(2018GXNSFAA281028))。

摘　　要：随着油气勘探观测环境愈发复杂,采集的地震数据常常掺杂各种噪声信号,导致勘探目标引起的有效微弱信号被覆盖,严重影响高精度的地震勘探数据解译,因而有效的压制地震勘探数据噪声显得越发重要。本文采用字典学习策略,将复杂地震数据进行分块,通过分块数据的字典学习获取字典原子,构建高精度的字典学习地震数据稀疏表示,通过两次迭代更新字典原子,进行数据去噪。将本文的字典学习算法应用于含随机噪声的模拟数据和实测地震勘探数据处理,验证该算法的可行性及有效性。结果表明,本文算法有效去除了随机噪声,保留了有效信号同相轴,提高了信噪比,可为复杂含噪地震数据的去噪处理提供新的技术手段。With the increasingly complex observation environment of oil and gas exploration, the seismic data collected are often mixed with various noise signals, resulting in the effective weak signal caused by the exploration target is covered, which seriously affects the high-precision seismic data interpretation, so it is more and more important to effectively suppress the seismic data noise. In this paper, the dictionary learning strategy is used to block the complex seismic data. The dictionary atoms are obtained through the dictionary learning of the block data, and the sparse representation of the seismic data is constructed by high-precision dictionary learning.The dictionary atoms are updated through two iterations for data denoising. The dictionary learning algorithm is applied to the processing of simulated and measured seismic data with random noise. The analysis results show that the algorithm can effectively removes the random noise while retains the effective signal phase axis, improves the signal-to-noise ratio which verifies the feasibility and effectiveness of the algorithm. The research results provide a new technical means for complex noisy seismic data denoising.

关 键 词：分块字典学习 地震数据去噪 迭代更新 实测数据处理 

分 类 号：O242[理学—计算数学] P315[理学—数学] P631[天文地球—地震学]