断层是控制油气田形成和分布的主要因素，查明断层的形态及分布是油气勘探的重要任务。传统的断层解释方法易受主观因素的影响，可靠性和精度不高，工作效率低。随着对断层描述方法的不断改进,使得断层的自动识别成为了可能。本文围绕断层自动识别技术中存在的主要问题，提出了用于断层线自动追踪的改进蚁群算法。本文主要内容分三个部分进行论述： 第一部分详细介绍了三代相干体算法、GST相干体算法和方差数据体算法，选用第三代相干体作为断层自动识别的断层属性体，为断层的自动追踪提供了依据。 第二部分简要阐述了蚁群算法的基本原理，重点讨论了基于图像特征的蚁群改进算法。考虑到断层识别与图像边缘检测的某些相似性，提出了蚁群算法在含噪声图像边缘检测中的应用原理和实现过程，并对该算法进行改进,实现了蚁群算法在断层线提取中的应用，通过实例讨论了算法中参数的影响，验证了算法的有效性。 第三部分首先介绍了断层线优化方法，包括数学形态学中的膨胀、腐蚀、开运算和闭运算，以及图像处理的细线化技术。然后选用闭运算和细线化方法对蚂蚁的追踪结果进行优化，得到连续性好、分辨率更高的断层线。 通过对某油田的实际资料进行处理，通过对水平切片和剖面进行断层线追踪，验证了本文方法的有效性。

The faults is the main factor that controls the formation and distribution of oil and gas fields, and it is the most important task to identify the morphology and distribution of faults in oil and gas exploration. Traditional method of fault interpretation is influenced by subjective factors, so that the reliability and accuracy is not high, and the efficiency of the work is low. With the continuous of improvement, the method of fault description makes it possible to identify faults automatically. According to the main problems existing in the fault automatic identification technology, this study proposed an improved ant colony algorithm for fault line automatic tracking .The main content is divided into three parts: The first part introduced three generations coherency cube algorithm, GST coherency cube algorithm and variance data cube algorithm, then adopted the improved third generation coherency cube as fault attributes cube of fault automatic identification, providing a basis for the later track of faults. The second part focuses on the improved ant colony algorithm based on images. Considering certain convergences of the fault identification and image edge detection, this paper proposed the application principle and the realization process of ant colony algorithm in the edge detection of the noise image, By improving the algorithm, and realized the application of ant colony algorithm to extract the fault line. Through the instance parameter in the algorithm is discussed, verified the effectiveness of the proposed algorithm. The third part introduced the optimization method of fault line, including the dilation, erosion, corrosion, opening operation and closing operation of mathematical morphology, also the graph thinning technology of image processing. The study optimizd the fault line of the ant tracking results through using the appropriate structural elements according to the tracking result, and could get binary used to extract fault. Through processing the actual oilfield data, verified the effectiveness of the method through discussing the tracking strategy according to the track of horizontal slice and vertical slice.