基于Curvelet变换的地震相干噪声去除

发布时间：2018-02-15 03:23

本文关键词： Curvelet变换形态成分分析(MCA) 地震去噪稀疏表示　出处：《哈尔滨工业大学》2017年硕士论文　论文类型：学位论文

【摘要】：地震勘探主要分为地震数据采集,地震数据处理,以及地震资料解释这三大环节。其中,地震数据采集指的是人为的通过放炮等方式激发地震波,然后利用检波器进行接收的过程。实际采集的过程中受到自然条件的限制,或者采集设备的影响等,采集到的数据往往会受到各种噪声的污染,严重影响后续的解释。因此,去噪这项工作迫在眉睫。根据噪声传播规律的不同,将噪声分为相干噪声(非随机噪声)和随机噪声这两大类。其中,相干噪声又可以分为线性噪声、多次波、面波等。本文的研究主要集中在对于地震相干噪声中的面波以及线性噪声的去除上。线性噪声的特点为:视速度稳定、且与有效信号的视速度不同;在地震记录上可以清楚地看出它们的同相轴倾角非常有规律;频率与有效反射波接近。Curvelet具有多尺度、多方向、紧框架、在时间域为近似线状的特征,可以最优的稀疏表示地震数据。不同频率、方向以及空间位置的地震波组在经过Curvelet变换后被很好地分离开。根据线性噪声与有效信号的视速度差异,空间位置差异等对含噪的地震数据进行Curvelet变换,分析Curvelet系数的特征差异,选择适当的阈值,对系数进行硬阈值处理,然后重构回去,取得了很好的去噪效果;并与传统f-k滤波去除线性噪声的方法进行对比,显示其在线性噪声压制方面的优势。由于形态成分的多样性,以及信号稀疏表示理论的发展,形态成分分析(MCA)理论被提出。该方法对于复杂信号的分离有着很好的效果,因此,被广泛应用于地震有效信号与噪声的分离当中。本文将此方法应用于面波压制中,取得了很好的效果。地震波在传播的过程中会有一部分沿着地表传播,并且它们之间相互影响叠加,这种波被称为面波。面波通常呈“扫帚状”,具有很强的能量、很低的频率、速度也很低、并且方向近乎垂直。考虑到面波的上述特点选择Curvelet变换作为面波的稀疏表示字典。考虑到地震数据的体波在形成过程中具有很强的局部特征,选择局部离散余弦变换对其进行稀疏表示。然后应用形态成分分析(MCA)法建立分离面波的模型,通过块松弛法进行求解,将面波与有效信号分开,从而达到压制面波的效果。将本方法应用于合成数据以及实际数据中,并与传统f-k滤波进行对比,显示其在面波压制上的优势。
[Abstract]:Seismic exploration is mainly divided into three parts: seismic data acquisition, seismic data processing, and seismic data interpretation. Among them, seismic data acquisition refers to the artificial firing of seismic waves. Then the process of receiving by using geophone. The process of actual acquisition is limited by natural conditions, or the influence of acquisition equipment, etc., the collected data will often be polluted by various noises, which will seriously affect the subsequent interpretation. According to the different laws of noise propagation, the noise is divided into two categories: coherent noise (non-random noise) and random noise. Among them, coherent noise can be divided into linear noise and multiple waves. Surface wave and so on. The research in this paper mainly focuses on the removal of surface wave and linear noise in seismic coherent noise. The characteristics of linear noise are that the apparent velocity is stable and different from the apparent velocity of the effective signal; From the seismic records, it can be clearly seen that their concentric inclination is very regular, and the frequency is close to the effective reflection wave. Curvelet has the characteristics of multi-scale, multi-direction, compact frame, which is approximately linear in time domain. Seismic wave groups with different frequencies, directions and spatial positions are well separated after Curvelet transform. According to the apparent velocity difference between linear noise and effective signals, The spatial position difference is used to transform the noisy seismic data into Curvelet transform, the characteristic difference of Curvelet coefficient is analyzed, the proper threshold value is selected, the coefficient is processed by hard threshold value, and then reconstructed back, and good denoising effect is obtained. Compared with the traditional f-k filtering method, it shows its superiority in linear noise suppression. Because of the diversity of morphological components and the development of signal sparse representation theory, The theory of morphological component analysis (MCA) has been put forward. This method has good effect on the separation of complex signals, so it is widely used in the separation of seismic effective signals and noise. In this paper, the method is applied to the suppression of surface waves. It works very well. Seismic waves propagate along the surface of the earth in part, and they interact with each other. These waves are called surface waves. Surface waves are usually "broomlike," with very strong energy and very low frequencies. Considering the above characteristics of surface wave, Curvelet transform is chosen as the sparse representation dictionary of surface wave. Considering that the volume wave of seismic data has very strong local characteristics during the formation process, The local discrete cosine transform is chosen to represent it sparsely, and then the model of the separated surface wave is established by using the morphological component analysis (MCA) method, and solved by the block relaxation method, the surface wave is separated from the effective signal. The method is applied to the synthetic data and the actual data, and compared with the traditional f-k filter, the advantages of the method in surface wave suppression are shown.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P631.4

【相似文献】