基于压缩感知的Dreamlet域数据重构方法及应用

Seismic data reconstruction in Dreamlet domain based on compressive sensing

随着压缩感知(CS)理论的完善,逐步发展形成了基于该理论的新的信号处理技术。近年来,在石油地震勘探领域,基于该理论的随机稀疏数据采样、数据重构及规则化和稀疏采样观测系统优化设计等方面的研究取得了重要进展。本文在Ru-Shan Wu博士等提出的Dreamlet域数据重构技术基础上,针对实际地震数据在时间和空间上剧烈变化以及存在较强干扰背景情况,通过优化重构参数和流程,对随机稀疏采样的模拟和真实地震数据进行了重构和对比分析。模拟和实际数据应用显示,该技术是一种高效和高质量的地震数据重构方法。

A new signal processing technique was developed recently due to the progress of compressive sensing theory,which is attracted attention in petroleum and nature gas exploration.Recently great progress has been made in seismic data random sparse sampling,seismic data reconstruction,seismic data regularization,and optimized geometry design.In this paper,we propose a new data reconstruction method based on the data reconstruction method in Dreamlet domain presented by Ru-Shan Wu.According to energy great variation in time and space domain and strong noise of seismic data,we reconstruct and compare synthetic data of random sparse sampling and real seismic data by optimizing parameters and the reconstruction process.Synthetic and real data tests show that the proposed method is a high-efficient and high-quality one.