二维多级中值滤波技术在随机噪声消除中的应用初探
|
摘要
二维多级中值滤波技术是信号处理中的一项成熟技术,它不仅具有自适应性,而且具有窗口的可变性。本文将这项技术用于地震勘探中的噪声压制,重点分析了噪声条件与滤波因子长度的关系。本文中定义的噪声条件包括噪声强度及噪声分布密度两个因素。当噪声条件固定时,二维多级中值滤波效果以5点滤波长度为最佳;当固定噪声密度时,改变噪声强度对滤波长度的选取不起决定性作用;当固定噪声强度,改变噪声密度对滤波长度有很大的影响。此法通过实际资料的处理,不仅能有效地压制随机噪声,而且在一定程度上削弱了浅层折射干扰
2D multi-level median filtering technique is a mature technique in signal processing,which has not only auto-adapted ability,but also has changeable feature of windows.The paper uses this technique to eliminate the noises in seismic exploration,and emphasizes analysis of the relationship between the noise conditions and the length of filtering factor.The noise conditions defined by the paper include two factors:noise strength and noise distribution density.The effects of 2D multi-level median filtering are better by selecting the filtering length as 5 points when the noise conditions are fixed;changing the noise strength has no determinate role for selection of filtering length when noise density is fixed;changing noise density has significant influence on filtering length when noise strength is fixed.It is shown by practical data processing that the method not only can effectively suppress random noises,but also can reduce the shallow refraction interference in some degree.
2D multi-level median filtering technique is a mature technique in signal processing,which has not only auto-adapted ability,but also has changeable feature of windows.The paper uses this technique to eliminate the noises in seismic exploration,and emphasizes analysis of the relationship between the noise conditions and the length of filtering factor.The noise conditions defined by the paper include two factors:noise strength and noise distribution density.The effects of 2D multi-level median filtering are better by selecting the filtering length as 5 points when the noise conditions are fixed;changing the noise strength has no determinate role for selection of filtering length when noise density is fixed;changing noise density has significant influence on filtering length when noise strength is fixed.It is shown by practical data processing that the method not only can effectively suppress random noises,but also can reduce the shallow refraction interference in some degree.
引文
[1] 宋焕生等.一种新的自适应多级中值滤波器.信号处理,1996,12(4)
[2] 宋焕生等.用于图像处理的自适应多级中值滤波器.西安交通大学学报,1998,32(1)
[3] NieminenA et al.A new class of detail- preservingfilters for image processing.IEEE TransonPAMI,1987,9(1)
[4] ArceG R andFosterR E.Detail- preserving ranked-order based filter for image processing.IEEE Tran-sonASSP,1987,37(1)
[5] L i et al.Relationship of the m ulti- shell to m ultistageand standard median filters.IEEE Trans onImageProcessing,1995,4(8)
[6] SunTong et al.Image restoration by multistage ada-ptive center weighted median filters.Proceedings ofSPIE-TheInternationalSociety forOpticalEngi-neering,1994,2180
[7] WangXin.Adaptive multistage m edian filter.IEEETrans onSignalProcessing,1992,40(4)
[8] 冯日亘等.中值滤波对信号相位和形状影响的研究.石油物探,2002,41(1)
[9] 刘洋等.加权中值滤波对信号相位影响的研究.长春科技大学学报,2001,(增)
[10] 钱晶洁,王欣.加权多级中值滤波.山东工业大学学报,2000,30(5)
[11] 邵敏等.二维中值滤波在电磁测深曲线去噪中的应用.物探化探计算技术,2001,23(3)
[12] 王欣.多级中值滤波的新算法.山东电子,1996,3
[2] 宋焕生等.用于图像处理的自适应多级中值滤波器.西安交通大学学报,1998,32(1)
[3] NieminenA et al.A new class of detail- preservingfilters for image processing.IEEE TransonPAMI,1987,9(1)
[4] ArceG R andFosterR E.Detail- preserving ranked-order based filter for image processing.IEEE Tran-sonASSP,1987,37(1)
[5] L i et al.Relationship of the m ulti- shell to m ultistageand standard median filters.IEEE Trans onImageProcessing,1995,4(8)
[6] SunTong et al.Image restoration by multistage ada-ptive center weighted median filters.Proceedings ofSPIE-TheInternationalSociety forOpticalEngi-neering,1994,2180
[7] WangXin.Adaptive multistage m edian filter.IEEETrans onSignalProcessing,1992,40(4)
[8] 冯日亘等.中值滤波对信号相位和形状影响的研究.石油物探,2002,41(1)
[9] 刘洋等.加权中值滤波对信号相位影响的研究.长春科技大学学报,2001,(增)
[10] 钱晶洁,王欣.加权多级中值滤波.山东工业大学学报,2000,30(5)
[11] 邵敏等.二维中值滤波在电磁测深曲线去噪中的应用.物探化探计算技术,2001,23(3)
[12] 王欣.多级中值滤波的新算法.山东电子,1996,3
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心