复值相干模量蚂蚁体技术

Ant tracking technology based on complex-valued coherence

史刘秀¹ 王静波² 张如伟³ 黄滨滨¹

中国石油大学（北京）地球物理与信息工程学院，北京 102249 中国石化勘探分公司勘探研究院，四川 成都 610041 广州海洋地质调查局，广东 广州 510760

College of Geophysics and Information Engineering, China University of Petroleum, Beijing 102249, China Research　Institute　of　SINOPEC　Exploration　Company,　Chengdu　610041,　China Guangzhou　Marine　Geological　Survey,　Guangzhou　510760, China

蚂蚁体算法是通过模拟自然界蚂蚁的觅食行为而总结出的一种正反馈仿生学算法机制。蚂蚁体技术已经广泛应用于断层预测，实现断层裂缝的自动追踪。文中回顾了蚂蚁体技术的发展历程，研究了影响蚂蚁体追踪质量的6个参数。传统的蚂蚁体技术基于地震方差体，信噪比和分辨率相对较低，文中提出了基于多道局部复值相干的蚂蚁体预测断层技术。该方法将复值相关算法和蚂蚁体算法结合起来追踪断层，得到复值相干模量蚂蚁体。实际应用表明，该方法既克服了多道局部复值相干边缘检测技术受地层边界影响大、对多组断层预测困难的缺点，又克服了传统蚂蚁追踪方法对断层预测不全面、干扰信息多的缺点。同时，人工解释结果验证了该方法具有较高的准确度和可靠性。

Ant algorithm is a kind of positive feedback mechanism and bionics algorithm, which is summarized by simulating foraging behavior of natural ants. The ant tracking technology has been widely used in fault prediction and automatic tracking of fracture. This paper reviews the research history of ant tracking technology and studies the effect of six parameters of ant tracking. Based on seismic variance, SNR（Signal to noise ratio） and resolution of traditional ant tracking technology is relatively low, so this paper proposes ant tracking technology by multichannel local complex-valued correlation for fault prediction. This method will get complex-valued coherence ant tracking data by combining complex-valued coherence algorithm and ant algorithm. Actual application shows that this method not only overcomes the disadvantage of the multichannel local complex-valued coherence edge detection technology affected largely by stratigraphic boundary and difficulty predicting multi-faults, but also overcomes the problem that traditional ant tracking method is not comprehensive and has some interference information for fault prediction. At the same time, the artificial interpretation results validate the high accuracy and reliability of the method.

10.6056/dkyqt201505001