中文摘要：

边缘检测是数字图像处理中基础而又重要的课题,Sobel算子是图像边缘检测中常用的方法之一。首先将Sobel与其他几种常用的边缘检测算子进行分析和比较,对经典的Sobel算子进行更深入的研究,Sobel算子的优点在于检测速度快,对噪声都具有平滑作用,对噪声具有一定的抑制能力;虽然产生了较好的边缘效果,但也存在一些问题：检测出了一些伪边缘,使得边缘比较粗,降低了检测定位精度。针对Sobel算子存在的问题,研究学者提出了一些改进的方法,如：扩展检测模板,边缘跟踪等方法,在分析这些方法的基础上,提出一种新的简单的Sobel算子改进方法。这种方法在经典的Sobel算法检测出的边缘的基础上对边缘进行细化,实验证明该方法有效解决了Sobel算子边缘检测时边缘过粗的问题,得到的边缘较细,定位精确。

英文摘要：

Edge detection is a fundamental and important issue in digital image processing. Sobel operator is one of the methods commonly used in edge detection.This paper starts with analysis and comparison between Sobel operator and several other commonly used edge detection methods,contributing to a more in-depth study on classical Sobel operator.Sobel operator＇s advantage is its faster detection rate,with unique function to smooth and restrain noise to certain extent.Although the edge effect is great,there are still some problems;some false edges are detected,which makes edges rougher and reduces accuracy of location detection.Researchers proposed some improved methods to solve the problems of Sobel operator,such as： detection template expansion,edge tracking,etc,.Based on the analysis of these methods,the paper presents a new and simple Sobel operator enhancement.This approach realizes edge refinement based on classic Sobel edge detection algorithm.Experimental results show that the method effectively solves the problem of rough edges in Sobel operator,achieves finer and precise positioning.