基于图像边缘检测方法提取地震剖面同相轴

地球物理勘探中，地震剖面中的同相轴对地质分层、断层解释、有无形成空洞以及其它异常体的检测中作用重要.根据地震剖面上同相轴具有的性质，采用图像边缘检测方法提取同相轴简单且有效.由于边缘检测中常用的Sobel算子和Canny算子检测同相轴时对噪声免疫差，常因噪声因素引起边缘信息缺失导致同相轴提取不连续，本文采用中值滤波并对Sobel算子进行改进，对地震剖面图进行预处理，结合均方差自动确定阈值并进行异常检测，提取完整同相轴.实验结果表明应用该方法可以提取连续、准确的同相轴.

Extraction of cophasal axes on seismic ections based on the edge detection method

Cophasal axes on seismic sections play an important role in geologic stratification, fault interpretation, detection of cavities and other anomalous bodies in geophysical prospecting. According to the characteristics of cophasal axes, it is easy and effective to use the edge detection method to extract cophasal axes. The ordinary edge detection methods such as Sobel and Canny operators are immune to noise, which result in cophasal axes discontinuity due to lack of edge information caused by noise. In order to extract cophasal axes more continuous and immune to noise, a novel approach is developed based on edge detection. Firstly, the median filter and improved Sobel method are used to preprocess seismic section images. Secondly, the threshold value is determined automatically based on variance to detect abnormal information. Finally, the complete cophasal axes are extracted after a post-processing. Experimental results show that the proposed method is effective in extracting continuous and correct cophasal axes.