基于双密度双树复小波变换的地震信号噪声压制方法研究

发布时间：2018-03-02 05:37

本文关键词： 双密度双树复小波变换双变量模型随机噪声小波系数最大后验概率估计　出处：《合肥工业大学》2017年硕士论文　论文类型：学位论文

【摘要】：压制地震信号中的随机噪声,提取有效信息是地震信号处理中的重要环节之一,噪声压制的好坏直接影响后面的地质解释工作。由于双密度双树复小波变换具有多方向选择性以及近似平移不变性等优点,所以本文在双密度双树复小波域建立统计模型压制地震信号中的随机噪声。本文主要研究内容如下:1、综述了地震信号降噪的研究背景、研究意义以及国内外研究现状,介绍了噪声的种类特点以及现阶段地震信号降噪的评价标准,并给出了相关的定量计算公式。2、介绍了离散小波的相关理论知识;其次针对离散小波的不足之处,详细地阐述了两种具有更多方向选择性的小波:双树复小波和双密度小波;最后基于双树复小波和双密度小波优良特性的基础上,着重介绍具有更多方向选择性以及近似平移不变性等优点的双密度双树复小波。并通过滤波器组结构示意图进一步解释了几种小波变换的分解过程。3、建立双密度双树复小波域的实部和虚部小波系数间的双变量模型对地震信号中的随机噪声进行压制。地震信号经过双密度双树复小波变换后,同尺度和同方向的实部和虚部小波系数之间具有很强的相关性。因此,建立双密度双树复小波实部和虚部系数的双变量模型,然后利用最大后验概率估计方法进行估计得到源地震信号小波系数的估计值,最后对源地震信号小波系数的估计值进行双密度双树复小波逆变换重构得到降噪后的地震信号。为验证本文算法的有效性,分别对合成地震记录和实际地震信号进行降噪处理,实验结果表明本文方法能够有效地压制随机噪声并且能够保留较多的有效反射波信号,提高信噪比的同时具有较高的保真度。
[Abstract]:Suppress random noise in the seismic signal, effective information extraction is one of the important link of seismic signal processing, noise suppression will directly influence behind the geologic interpretation. Due to the double density dual tree complex wavelet transform with multi direction selectivity and approximate shift invariance etc., so the double density dual tree complex wavelet domain statistical model pressing the random noise in the seismic signal. The main research contents of this paper are as follows: 1, summarized the research background of seismic signal denoising, significance and research status at home and abroad research, introduces the types of noise characteristics and noise evaluation stage of seismic signal, and the formulas for calculating.2 gives quantitative, introduces relevant theoretical knowledge discrete wavelet; secondly aiming at disadvantages of discrete wavelet, elaborates two kinds of more selective dual tree complex wavelet direction: small Wave and double density wavelet; finally, based on the dual tree complex wavelet transform and double density wavelet excellent characteristics, emphatically introduces the direction selectivity and has more advantages of double density dual tree complex wavelet approximate shift invariance etc.. And through the filter bank structure diagram further explained the decomposition process of.3 wavelet, is used to suppress the random noise of two variables model of real and imaginary part of wavelet coefficients of double density dual tree complex wavelet domain between the seismic signal in seismic signal. After the double density dual tree complex wavelet transform, has a strong correlation between the real and imaginary part of wavelet coefficients of the same scale and the same direction. Therefore, the bivariate model established the double density dual tree the real and imaginary part of complex wavelet coefficients, and then use the method to estimate the maximum a posteriori estimate to estimate the source of seismic signal wavelet coefficient values at the end of the earthquake source Estimate number of wavelet coefficient value of seismic signal is the double density dual tree complex wavelet transform is reconstructed after noise reduction. In order to verify the effectiveness of the proposed algorithm, respectively on the synthetic seismogram and actual seismic signal de-noising, the effective reflection wave signal experimental results show that this method can effectively suppress noise and to retain more. To improve the signal-to-noise ratio and has high fidelity.

【学位授予单位】：合肥工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P631.4

【相似文献】