基于Canny算子改进的边缘检测算法  

Edge detection algorithm based on improved Canny operator

作　　者：王文豪[1] 姜明新[1] 赵文东[1] 

机构地区：[1]淮阴工学院计算机与软件工程学院,江苏淮安223003

出　　处：《中国科技论文》2017年第8期910-915,共6页China Sciencepaper

基　　金：国家自然科学基金资助项目(61403060);淮安市科技项目(HAR2015007)

摘　　要：针对经典Canny算子在噪声去除、高低阈值选取和边缘断裂等问题,提出了1种改进的Canny算子边缘检测算法。首先用自适应中值滤波取代高斯滤波;然后采用改进的Otsu算法根据图像梯度幅值自适应地生成Canny算子的高、低双阈值;最后采用多特征融合技术修复断裂的边缘。实验结果表明,改进的算法具有较好的检测精度和抗噪声能力,能有效地修复图像断裂边缘,很好地控制断裂边缘的错误连接。To solve the problem of the noise removal, high-low threshold and edge breaking in the traditional canny algorithm, an edge detection algorithm is proposed based on improved Canny algorithm. Firstly, noise is removed with the adaptive median fil-ter instead of Gaussian filter. Then, the high-low threshold of canny algorithm is adaptive generated by the improved Otsu algo-rithm based on image gradient magnitude. Finally, multi-feature fusion technique is used to repair the broken edge. The experi-mental results show that the improved algorithm not only has better detection precision and the ability of anti-noise in the edge de-tection, but also can effectively repair the broken edge and control the errors of the edge connection.

关 键 词：CANNY算法 边缘检测 高斯算子 自适应中值滤波 改进的Otsu算法 边缘修复 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]