基于改进迭代去噪法的多震源地震数据分离  

Seismic data separation of simultaneous source based on an improved iterative denoising method

作　　者：郭建宏 成景旺 陈芷若 杨兵 严皓 GUO Jian-Hong;CHENG Jing-Wang;CHEN Zhi-Ruo;YANG Bing;YAN Hao(Geophysics and Oil Resource Institute,Yangtze University,Wuhan430100,China;Key Laboratory of Exploration Technologies for Oil and GasResources,Ministry of Education,Yangtze University,Wuhan 430100,China;College ofMining Engineering,Taiyuan University of Technology,Taiyuan 030024,China;TianjinBranch of CNOOC Ltd.,Tianjin 300452,China)

机构地区：[1]长江大学地球物理与石油资源学院,湖北武汉430100 [2]长江大学油气资源与勘探技术教育部重点实验室,湖北武汉430100 [3]太原理工大学矿业工程学院,山西太原030024 [4]中海石油有限公司天津分公司,天津300452

出　　处：《物探与化探》2019年第5期1054-1063,共10页Geophysical and Geochemical Exploration

基　　金：国家自然科学基金(41504102);湖北高校省级大学生创新训练项目(201810489075)

摘　　要：多震源激发技术与单震源采集方法相比,极大地提高了地震勘探效率,其混合数据的分离是影响成像的关键因素。将多震源混合数据的伪分离记录从共炮道集分选到其他道集时,次震源地震信号表现为随机噪声,可通过迭代去噪的方法进行分离。本文对传统的迭代预测去噪法进行改进,在每次去噪前加入上一次分离结果来增加迭代去噪的稳定性,并结合多方向矢量中值滤波来进行随机噪声的去除。通过理论模型的多震源混合数据分离试验,表明本文提出的方法具有更好的分离效果,且更为强健,在不同滤波参数下均能保持一定的稳定性。将该方法应用于实际海洋混合地震数据的分离,同样能得到较好的分离结果,具有实际应用价值。Compared with single source acquisition method,simultaneous source acquisition technology can greatly increase field acquisition efficiency,and the separation of blended data is the key.When pseudo-segregated records of mixed simultaneous source data are sorted from common gathers to other gathers,the seismic signals of secondary sources show random noise,which can be separated by iterative denoising method.In this paper,the traditional iterative prediction denoising method is improved by adding the last separation result before each denoising to increase the stability of the iterative denoising,and combining with Multidirectional Vector Median Filtering to remove random noise.Through the separation test of simultaneous source mixed data of theoretical model,it is shown that the proposed method has better separation effect and is more robust,and can be operated under the condition of different parameters and maintain certain stability.Finally,the method is applied to the separation of Simultaneous Source Marine Line data,and a better separation result can be obtained.The method has practical application value.

关 键 词：多震源激发 混合数据分离 迭代预测相减法 多方向矢量中值滤波 

分 类 号：P631.4[天文地球—地质矿产勘探]