摘要
为提高Canny算子在噪声环境下的鲁棒性和图像边缘提取精度,提出了一种改进的Canny算子。所提算子采用自适应高斯滤波器替代传统高斯滤波器,通过非线性插值法改进非极大值抑制过程,细化图像边缘的宽度和精度。采用改进的Otsu算子和区域生长法动态调整边缘检测中的高低阈值,降低固定阈值带来的误判、漏检。将所提算子和Canny算子、Sobel算子以及改进Canny-Devernay亚像素边缘检测算子进行对比,结果表明所提出的改进Canny算子边缘定位误差降低至1.21 px,F 1值达到0.94,且比传统Canny算子的运行时间仅增加0.01 s,对工业检测、医学影像等有较高要求的边缘提取场景具有一定的实用价值。
To improve the robustness of the Canny algorithm in noisy environments and the accuracy of image edge extraction,an improved Canny algorithm was proposed.The proposed algorithm replaced the traditional Gaussian filter with an adaptive Gaussian filter,and improved the non-maximum suppression process via a nonlinear interpolation method to refine the width and accuracy of image edges.In addition,an improved Otsu algorithm and region growing algorithm were employed to dynamically adjust the high and low thresholds in edge detection,reducing false detections and missed detections caused by fixed thresholds.The proposed algorithm was compared with the Canny algorithm,Sobel algorithm and the improved Canny-Devernay subpixel edge detection algorithm,and the results showed that the edge localization error of the improved Canny algorithm is reduced to 1.21 px,the F 1 value reaches 0.94,and the running time is only 0.01 s longer than that of the traditional Canny algorithm.It has certain practical value for edge extraction scenarios with high requirements such as industrial inspection and medical imaging.
作者
江铁军
李春峰
JIANG Tiejun;LI Chunfeng(School of Broadcasting,Hosting and Film&Television,Anhui Broadcasting Movie and Television College,Hefei 230011,China;School of Computer Science,Jiangsu University of Science and Technology,Zhenjiang 212100,China)
出处
《成都工业学院学报》
2025年第6期28-33,共6页
Journal of Chengdu Technological University
基金
安徽省重大教学研究项目(2020jyxm0279)。
