基于奇异值分解的废道自动识别算法  

Automatic rejection algorithm of abnormal seismic traces via singular value decomposition

作　　者：曹永生[1] 陈金焕[1] 王小青[1] 毕进娜 林庆富[1] 

机构地区：[1]中国石油化工股份有限公司石油物探技术研究院,南京211103

出　　处：《地球物理学进展》2015年第5期2120-2124,共5页Progress in Geophysics

摘　　要：在地震资料处理中,废道检测是一项基础性工作,对后续处理进度和质量有较大影响.奇异值分解技术广泛应用于地震资料处理,本文首次将奇异值分解应用于废道识别中,给出一种基于奇异值分解的废道自动识别算法,通过提取平均振幅、主频、过零点个数、相关系数和衰减因子等5个地震道属性特征,利用奇异值分解技术找出正常道所具有的属性特征向量,根据向量距离找出废道,最后通过实际生产所用地震资料进行了测试验证,结果表明,本算法可以快速准确地识别废道.The detection of abnormal seismic traces is a basic work in seismic data processing. And the result has a significant impact on subsequent processing and quality.Today,the singular value decomposition（SVD）technique has been used in seismic data processing widely.This paper applies SVD to identify abnormal seismic traces firstly and gives an automatic rejection algorithm of abnormal seismic traces via SVD.Firstly,the seismic trace attribute features, which contain average amplitude, main frequency,zero crossing number,the correlation coefficient and attenuation factor,were extracted.Secondly,the SVD technique was used to find the property feature vector of the normal trace. Then,the abnormal traces were identified by the distance of vectors.Finally,the algorithm was tested for the real seismic data through the oriented developed software system.The results show that this algorithm can identify the abnormal seismic traces quickly and accurately.

关 键 词：废道 奇异值分解 自动识别 道属性特征 向量距离 

分 类 号：P631[天文地球—地质矿产勘探]