盆地深层火山岩位场分离与边界识别方法概述

A Review of Potential Field Separation and Boundary Identification Methods of Basin Deep Volcanic Rocks

当今盆地深层火山岩已成为油气勘探的热点领域,通过位场分离和边界检测获得火山岩的重磁异常和分布情况,是火山岩油气勘探的首要环节。本文首先基于各种位场分离方法的分离能力,将它们归于二分法和多分法2类,并对能够胜任盆地深部火山岩重磁异常提取的多分位场方法的起源、发展和现状进行了阐述。分析认为匹配滤波受垂向上不同场源可能的频谱重叠影响,位场分离效果欠佳;小波分析通过小波多尺度分解和功率谱分析,可有效分离不同深度层次的位场并确定场源埋深;由于切割半径和异常体中心埋深具有等量关系,插值切割法可实现明确地质意义上的位场分层分离。因此,相较于匹配滤波,小波多尺度分析和插值切割法具有更好的位场分离效果。其次,对局部异常的边界检测方法进行了分类对比,认为直接导数计算类方法可有效确定浅部场源边界,但对于深部场源边界反应不灵敏。导数比值计算类方法可以将深、浅部场源引起的异常归一到同一水准,实现深、浅部场源边界的准确识别,因此,导数比值计算类方法是盆地深部火山岩边界检测的优先选择。此外,讨论了导数比值计算类方法的解析奇点和导数计算时引入的噪声对边界识别精度的影响及解决方案。

Currently,deep volcanic rocks in the basin have become a hot field for oil and gas exploration.It is crucial for volcanic oil and gas exploration to obtain the gravity and magnetic anomalies caused by volcanic rocks and their distribution through potential field separation and boundary identification.In this review,all types of potential field separation methods are first classified into two categories based on their separation ability:dichotomous and multi-division methods.The origin,development,and current situation of multi-division potential field methods,capable of extracting gravity and magnetic anomalies of basin deep volcanic rocks,are discussed in detail.Consequently,the matched filtering method does not have the desired effect of potential separation due to probable spectral overlap among different field sources in the vertical direction.The wavelet analysis method can effectively separate the potential field of field sources with different buried depths and determine their depths using wavelet multiscale decomposition and power spectrum analysis.The interpolation cutting method can realize the stratified separation of potential fields with specific geological significance due to the equivalent relationship between the cutting radius and the central depth of the abnormal body.Therefore,wavelet multiscale analysis and interpolation cutting methods are more effective for potential field separation than matched filtering.Second,boundary detection methods for local anomalies were classified and compared.Direct derivative calculation methods are considered effective in determining the boundary of shallow field sources but are not sensitive to those of deep field sources.Derivative ratio calculation methods can normalize anomalies caused by deep and shallow field sources to the same level,facilitating accurate boundary identification for both deep and shallow field sources.Therefore,derivative ratio calculation methods are preferred for detecting deep volcanic rock boundaries in basins.Additionally,the influences of the analytic singularity of derivative ratio calculation methods and the noise introduced in derivative calculation on the accuracy of boundary recognition and solutions are discussed.