摘要
近年来新兴的多种地震数据处理和解释技术都需要倾角作为先验信息,传统的倾角估计方法大都依赖成像剖面,这些方法不可避免地会受到成像质量的影响,而且空间倾角的三维估计更是依赖于多条线形成的成像体.二元倾角道集因其椭圆形反射波位置对三维倾角具有指示性的特点,可以用作对倾角的估计,这样不仅能避免低质量成像剖面带来的影响,同时也能够实现仅依赖一条成像线道集的三维倾角估计.然而该方法会消耗大量人力和时间并且其结果也依赖处理人员经验.二元倾角道集在一个工区中数量庞大,可以为数据驱动特征提取的深度学习算法提供样本支持,因此本文在二元倾角道集的基础上引入深度学习算法,提出一种基于改进二元倾角道集和VGG神经网络的倾角提取技术,实现倾角的自动估计,并将倾角应用于菲涅尔带相关孔径的估计中,最后通过某工区实际数据验证对倾角估计的准确性和网络的泛化性能,再将孔径用于稳相偏移来验证倾角在实际数据中应用的有效性.
Various emerging seismic data processing and interpretating techniques require dip angle of the strata as a priori information.Most traditional dip angle estimation methods rely on imaging profiles,which are inevitably affected by imaging quality.Moreover,the three-dimensional estimation of spatial dip angle relys on the image cube formed by multiple imaging sections.The 2D dip gathers can be used to estimate the dip because the position of the elliptical reflection is indicative of the 3D dip,which not only avoids the influence of low-quality imaging sections,but also realizes dip estimation on a single imaging line.However,this method relies on personal experience and consumes excessive manpower and time.Dip gathers in a seismic prospecting can provide sufficient sample support for data-driven deep learning algorithms to extract complex feature.Therefore,by introducing deep learning algorithm on the basis of 2D dip gathers.So by introducing deep learning algorithm,we come up with a dip angle automatic pickup method on the basis of 2D dip angle gathers and VGG network,and then apply the result in the estimation of Fresnel-zone-related aperture.Finally,the accuracy of dip angle estimation and the generalization performance of the network are verified by the practical data from an oil field,and the calculated aperture is then applied in stationary phase migration to verify the effectiveness of the application of the predicted dip angle in actual data.
作者
王晨源
陈娟
张雅雯
张江杰
WANG ChenYuan;CHEN Juan;ZHANG YaWen;ZHANG JiangJie(Key Laboratory of Petroleum Resources Research,Institute of Geology and Geophysics,Chinese Academy of Sciences,Beijing 100029,China;Innovation Academy for Earth Science,Chinese Academy of Sciences,Beijing 100029,China;University of Chinese Academy of Sciences,Beijing 100049,China;Exploration and Development Research Institute of PetroChina Changqing Oilfield Company,Xi'an 710018,China)
出处
《地球物理学报》
SCIE
EI
CAS
CSCD
北大核心
2023年第8期3403-3412,共10页
Chinese Journal of Geophysics
基金
国家自然科学基金(42074158,42030802)资助。
