小波包相关阀值去噪

Noise elimination with correlative threshold plus wavelet packet transformation

1. 成都理工大学信息工程学院,四川成都610059;2. 中油集团物探局,河北涿州072751

1. Information Engineering Institute of Chengdu University of Science and Technology, Chengdu 610059, China;2. Bureau of Geophysical Prospecting, China Petroleum and Natural Gas Group Corporation, Zhuozhou 072751, China

在地震勘探中,随机噪音是一种频带较宽,严重影响有效波信噪比的干扰波,一般的一维去噪方法效果都不理想。小波包变换是一种时频分析的方法,在分析中、高频方面优于小波变换,它结合相关阀值,能有效地在地震勘探剖面中消除随机干扰。分析了小波包优于小波变换的基本原理,以及小波包在地震记录中的分解、相关阀值去噪和地震记录重构,进行了实例分析对比,说明了小波包相关阀值法去噪优于其它去噪方法,且该方法计算速度快,简便可行。

Random noises are one of interference waves of wide waveband in seismic exploration. It has a strong impact on the signal to noise ratio of significant waves. Conventional 1 D noise eliminationmethods fail to yield desired results. Wavelet packet transformation is amethod for analysis of relations between time and frequency. It ismuch better than wavelet transformation in analyzing signals ofmedium and high frequencies. Along with correlative threshold, it can effectively eliminate random noises. This paper presents the principle of wavelet packet transformation, the decomposition of seismic records with wavelet packet transform, correlative threshold noise elimination and reconstruction of seismic records. Real data processing shows that noise elimination with correlative threshold plus wavelet pact transformation is superior to conventional 1-D noise eliminationmethods. Themethod is fast and easy to use.