本课题针对时空轨线型TFPF消噪模型中最优滤波轨线的选取问题，展开对地震勘探资料中同相轴分布形态检测算法的研究，进而设计地震勘探时空轨线TFPF消噪软件系统及其可视化操作界面。整个研究分别从确定型同相轴和不规则同相轴两种情况进行，对于直线型（或曲线型）的确定型同相轴拟采用Hough变换及其衍生算法实现同相轴的线条化处理；而对于不规则同相轴，拟采用整体近似法估计其整体变换趋势或分段法进行局部形态检测，进而确定合适的最优滤波轨线，为时空轨线TFPF消噪模型的建立及其去噪性能的改善提供重要依据。该课题的研究有助于促进地震勘探信号处理及噪声压制等理论及算法的发展。

This research discusses the morphological edge detection algorithms about the reflection events in seismic records, which is to applied to the optimal filtering trace selection in the spatio-trace time-frequency peak filtering(TFPF) random noise attenuation model. And then a set of software system about this model is to designed accordingly. The whole research consists of two main aspects in the edge detection: definite events and irregular ones. It aims to adapt the Hough transform and its derivative algorithms for the definite events and global approximation methods for the irregular ones. The obtained optimal filtering traces provides important basis for the spatio-trace TFPF random noise attenuation model and improve its performance greatly. This research will help promote the development of seismic signal processing and noise suppression theory and algorithms.