基于二维经验模态分解的位场数据分析及应用  

Analysis of potential field data and its application based on bidimensional empirical mode decomposition

作　　者：朱振宇[1] 刘国峰[1] 

机构地区：[1]中国地质大学(北京)地球物理与信息技术学院,北京100083

出　　处：《地球物理学进展》2016年第2期882-892,共11页Progress in Geophysics

基　　金：国家高技术研究发展(863)计划(2013AA063905-04)资助

摘　　要：二维经验模态分解(BEMD)是一种信号的多尺度分析技术,该方法可以对非线性、非平稳信号进行分析而且具有自适应性.本文实现了BEMD分解算法并对其进行了如下改进研究:针对数据特点,将边缘数据点视为极值点来解决边缘效应;通过对比,采用了径向基函数作为包络面重建方法,减少了计算误差;研究了不同筛分停止准则对分解结果的影响.本文对改进BEMD方法进行了以下应用:对长江中下游重力数据进行了多尺度分解,并根据分解的结果和现有地质认识,划定了部分主要断层分布;根据位场功率谱理论,对不同尺度的分解结果进行了场源深度估算,使BEMD分解结果有了对应的深度概念;应用本文方法对闽西南航磁数据进行了去工业噪声处理.Bidimensional empirical mode decomposition（ BEMD）is a multi-scale analysis of a signal, the method can analyze nonlinear and non-stationary signals and is self-adaptability. this paper implements BEMD algorithm and made the following improvements and study： to suit the data characteristics,the edge data points are deemed extreme points to resolve edge effects; by contrast,we use a radial basis function as an envelope surface reconstruction method to reduce the calculation error; we also study the effects of different screening criteria to stop decomposition results. In this paper,the improved BEMD method is applied two aspects： Firstly,the Yangtze River gravity data were divided into multiscale decomposition, and in accordance with the result of decomposition and existing geological knowledge we delineate some of the main fault,while using the power spectral theory to estimate source depth of different decomposition result, so BEMD decomposition results have corresponding depth concept; the second is the application of this method to western Fujian Southern magnetic data were processed to industrial noise.

关 键 词：重力异常分离 二维经验模态分解 径向基函数 去噪 

分 类 号：P631[天文地球—地质矿产勘探]