摘要
传统Canny边缘检测算法的效果易受噪声、光照等因素的影响,且需人工选择高斯滤波方差与滞后阈值,自适应性差.提出改进的Canny边缘检测算法,该算法采用复合形态滤波取代高斯滤波,降低噪声的影响,保持边缘强度与细节;采用加入了45°和135°方向模板的Sobel算子计算梯度幅值,提高边缘定位效果;根据图像的灰度均值和方差选取阈值,边缘检测更连续.实验表明:改进的算法比传统Canny算法具有更好的检测效果和抗噪性能.
The traditional Canny algorithm without the adaptive ability in the variance of Gaussion filtering and threshold,is susceptibale to noise,light and so on. This paper proposes an improved edge detection method based on Canny alogorithm. Firstly,compound morphology smoothing replaces Gaussian filtering,not only reduces the influence of noise,but also maintains the edge strength and details. Then,the gradient amplitude is calculated by using the Sobel detector with adding 450 and 1350 direction templates,so as to improve the effect of edge location. Finally,the thresholds is automatically seleced according to the mean of variance and gray-scale,so that the edge detection is more continuous. Experiment shows the improved algorithm has a good effect and anti-noise performance.
引文
[1]张震,马驷良,张忠波,等.一种改进的基于Canny算子的图像边缘提取算法[J].吉林大学学报(理学版),2007,45(2):244-248.
[2]薛丽霞,李涛,王佐成.一种自适应的Canny边缘检测算法[J].计算机应用研究,2017,27(9):3588-3590.
[3]wang Bing,Fan Shao-sheng.An improved Canny edge detection algorithm[C].2010 Second International WorkShop on Computer Science and Engineering,2009:497-500.
[4]李俊山,马颖,赵方舟,等.改进的Canny图像边缘检测算法[J].光子学报,2011,40(S1):50-54.
[5]ZHAO Hui-li,Qin Guo-feng,Wang Xing-jian.Improvement of Canny algorithm base on pavement edge detection[C].2010 3rd International Congress on Image and Singnal Processing,2010:964-967.
[6]付纪刚.基于支持向量机的边缘检测算法研究[D].西安:西安科技大学,2011.
[7]葛阿雷,史伟.改进的自适应形态学边缘检测[J].宁夏大学学报,2016,37(1):34-38.
[8]卞桂平,秦益霖.基于Canny算法的自适应边缘检测算法[J].电子设计工程,2017,25(10):53-56,60.
[9]吴翔,于微波,马艳辉,等.一种改进Canny图像边缘检测算法[J].影像科学与光化学,2016,34(1):161-121.