| 一种改进的Sobel算子边缘检测及细化算法 |
| |
| 引用本文： | 沈德海,鄂旭,侯建.一种改进的Sobel算子边缘检测及细化算法[J].渤海大学学报(自然科学版),2014(3):256-260. |
| |
| 作者姓名： | 沈德海 鄂旭 侯建 |
| |
| 作者单位： | 渤海大学 信息科学与技术学院,辽宁 锦州,121013 |
| |
| 基金项目： | 辽宁省高等学校实验室项目(No L2012397)；博士后基金项目(No2012M520158)；辽宁省“百千万人才工程”资助项目(No2012921058)；辽宁省教育厅项目(NoL2012400)． |
| |
| 摘 要： | 针对经典的Sobel算子存在的边缘定位精度不高和边缘提取较粗等缺点,对经典Sobel算法进行了改进:在原有的两个方向模板基础上增加了135°和45°2个方向模板,并通过非极小值抑制和邻域标准差叠加获取梯度图像,提高了边缘定位精度和增强边缘强度;对梯度图像在3×3邻域内采用梯度阈值结合边缘方向进行了边缘细化处理;实验证明,该算法不仅有效地解决了Sobel算法提取边缘过粗及定位不精确的问题,而且使图像边缘更连续、清晰.
|
| 关 键 词： | 边缘检测 Sobel算子 邻域标准差 边缘细化 |
An edge detection refinement algorithm based on improved sobel operator |
| |
| SHEN De-hai,E Xu,HOU Jian.An edge detection refinement algorithm based on improved sobel operator[J].Journal of Bohai University:Natural Science Edition,2014(3):256-260. |
| |
| Authors: | SHEN De-hai E Xu HOU Jian |
| |
| Institution: | (College of Information Science and Technology, Bahai University, Jinzhou 121013, China) |
| |
| Abstract: | Abstract： In order to solve the shortcomings of low edge position accuracy and coarser edge of the traditional Sobel Operator, an improved algorithm was proposed, the algorithm increases 135° and 45° directions templates based on the original two directions templates, obtains the gradient image through Non - Minimum Suppression and local standard deviation to improve the edge positioning accuracy and enhance edge strength. Finally com- bines the gradient image in 3 × 3 neighborhood with a gradient threshold combined edge direction. Experiments show that the algorithm not only solves the problem of the Sobel, but also obtains more consistent and crisp ed-ges. |
| |
| Keywords: | edge detection sobel operator local standard deviation edge refinement |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
