应用小波多尺度分析分离海洋重力异常场

发布时间：2018-03-05 19:56

本文选题：海洋重力异常场划分　切入点：频率域　出处：《中国海洋大学》2014年硕士论文　论文类型：学位论文

【摘要】：研究区域地质构造需要大量的地球物理信息，在众多地球物理勘探方法中，重力勘探以其易实现、快捷、适用性强的特点得到广泛应用，重力资料也是地球物理资料中范围最广，信息最全的资料，故重力资料在研究中往往起着基础性的作用，在大尺度、大深度的地质构造研究中，重力资料更是作为主要研究资料，重力场数据反演能提供其他物探资料所不能提供的信息。重力异常反映的是地下剩余质量的分布情况，主要是由地质构造或矿藏分布等引起的地下物质密度不均匀分布造成的，将重力异常数据反演就能得到地下地质构造信息，但是重力异常是地下剩余质量的综合反映，想要得到更加具体和详尽的构造信息，在反演前首先要将重力异常进行划分，将不同地质构造引起的异常分开，分开后的重力异常反演就能得到更具体的构造信息。传统的二分法是将重力异常划分为大尺度构造引起的区域异常和小尺度构造或地质体引起的局部异常，传统的划分方法有高次求导、解析延拓、趋势分析等，这些方法都是在空间域实现重力异常的划分，实际上区域异常与局部异常在频率上一般存在较明显的差异，我们可以在频率域上根据异常的频率差异实现异常的划分，小波变换是一种较完美空间域与频率域转换工具，具有多尺度分析的优点。并且通过Mallat算法与滤波器组的结合可以快速实现，小波的多尺度分解也更符合反演问题多解性的需求。小波分析的上述优点使得它理论上能够有效的实现重力异常的划分。本文主要是对小波多尺度分析实现重力异常划分的技术研究，并通过实际海洋重力资料的处理验证小波多尺度分析法的实际应用效果，结合其他物探资料对划分结果给出相应的构造分析。论文的主要研究内容和路线为：首先研究了重力位场和重力异常的基本特征与原理，研究了解析延拓法和高次求导实现异常划分的数学原理并编程实现，之后研究了解了小波变换多尺度分析原理并结合Matlab编程实现，研究掌握了小波母函数选取原则，随后建立重力异常数据模型，，分别使用解析延拓、高次求导和小波多尺度分析实现异常划分，比较三种方法的优缺点，在小波分析时尝试不同的母小波和不同的分解阶次，验证了小波母函数的选取原则，选取了最符合实际情况的分解阶次，最后将南海和东海某区域的海洋重力异常卫星资料进行了小波多尺度分析，结合相关性分析及地形数据对小波分解结果进行了构造分析。研究表明小波分析能够快速、高效、准确的实现重力异常的划分，和解析延拓法及高次求导法相比有明显优势，实际海洋重力异常资料处理结果表明，小波分解的四阶逼近值主要是由莫霍面起伏引起的，而四阶细节则主要反映了沉积基底面的起伏。
[Abstract]:The study of regional geological structure requires a lot of geophysical information. Among many geophysical exploration methods, gravity exploration is widely used because of its easy realization, fast and strong applicability. Gravity data is also the most extensive and complete geophysical data, so gravity data often play a fundamental role in the study, in the large-scale, deep geological structure research, Gravity data as the main research data, gravity field data inversion can provide other geophysical data can not provide information. Gravity anomaly reflects the distribution of underground residual mass, which is mainly caused by the uneven distribution of underground material density caused by geological structure or mineral deposit distribution. The information of underground geological structure can be obtained by inversion of gravity anomaly data. But gravity anomaly is a comprehensive reflection of underground residual mass. In order to obtain more specific and detailed structural information, the gravity anomaly must be divided before inversion, and the anomalies caused by different geological structures should be separated. The conventional dichotomy is to divide gravity anomalies into regional anomalies caused by large scale structures and local anomalies caused by small scale structures or geological bodies. The traditional partitioning methods include high-order derivation, analytic continuation, trend analysis, etc. These methods all realize the division of gravity anomalies in spatial domain. In fact, there are obvious differences in frequency between regional anomalies and local anomalies. We can divide the anomaly in frequency domain according to the frequency difference of anomaly. Wavelet transform is a perfect conversion tool between spatial domain and frequency domain. It has the advantage of multi-scale analysis, and can be realized quickly by the combination of Mallat algorithm and filter bank. The multi-scale decomposition of wavelet is also more in line with the need of multi-solution of inverse problem. The advantages of wavelet analysis make it effective to partition gravity anomalies theoretically. In this paper, the technique of realizing gravity anomaly division by wavelet multi-scale analysis is studied, and the practical application effect of wavelet multi-scale analysis is verified by processing the actual marine gravity data. Combined with other geophysical data, the corresponding structural analysis of the division results is given. The main contents and routes of this paper are as follows: firstly, the basic characteristics and principles of gravity potential field and gravity anomaly are studied, and the mathematical principle of dividing anomaly by analytic continuation method and high order derivation is studied and realized by programming. Then, the principle of wavelet transform multi-scale analysis is studied and realized with Matlab programming, and the principle of selecting wavelet mother function is studied. Then, the gravity anomaly data model is established, and the analytic continuation is used respectively. High order derivation and wavelet multiscale analysis are used to realize abnormal partition, and the advantages and disadvantages of the three methods are compared. In wavelet analysis, different mother wavelets and different decomposition orders are tried, which verifies the selection principle of wavelet generating function. Finally, the wavelet multiscale analysis of the satellite data of marine gravity anomaly in the South China Sea and the East China Sea is carried out, and the wavelet decomposition results are constructed by combining the correlation analysis and topographic data. The research shows that wavelet analysis can realize the division of gravity anomaly quickly, efficiently and accurately, and has obvious advantages compared with analytic continuation method and high-order derivation method. The processing results of actual marine gravity anomaly data show that, The fourth order approximation value of wavelet decomposition is mainly caused by Moho surface fluctuation, while the fourth order detail mainly reflects the fluctuation of sedimentary basement surface.
【学位授予单位】：中国海洋大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：P631.1;P738.2

【参考文献】