摘要
Canny算子在边缘检测领域被广泛应用,但传统的Canny算子需要人为设定梯度阈值和合适的高斯标准差,因此自动化程度不高。为解决Canny算子的自适应性,发展了包括Otsu算法在内的自适应算法。但传统的Otsu算法存在定位精度不高,易受高斯滤波标准差影响等问题。在Otsu算法的基础上,通过对高斯滤波后梯度图像的分析与研究,对最大类间方差进行改进,从而提高自适应算法的稳定性。将改进后的类间方差与类内方差的比值作为阈值选取的评价标准,从而有效提高了自适应算法的定位精度。仿真表明,所提出的改进算法能有效提高Canny算法的定位精度和鲁棒性。
Canny operator is widely used in the field of edge detection,but it has low automation degree because the manmade setting of gradient threshold and suitable Gauss standard deviation. In order to improve the adaptivity of Canny operator,the adaptive algorithm including Otsu algorithm was developed. Unfortunately,the traditional Otsu algorithm has low positional accuracy,and is easy to be influenced by Gauss standard deviation. On the basis of studying the Otsu algorithm,the gradient image after Gaussian filtering is analyzed and researched,and the maximum interclass variance is improved to enhance the stability of adaptive algorithm. The ratio of the improved interclass variance and intraclass variance is taken as the evaluation criterion of threshold selection to improve the positional accuracy of adaptive algorithm. The simulation results show that the improved algorithm proposed in this paper can improve the positioning accuracy and robustness of Canny algorithm effectively.
引文
[1]刘芳,邓志仁.基于自适应小波阈值和双边滤波器的去噪算法[J].系统仿真学报,2014,26(12):2934-2938.LIU Fang,DENG Zhiren. Denoising algorithm based on adaptive wavelet threshold and bilateral filter[J]. Journal of system simulation,2014,26(12):2934-2938.
[2]戴青云,余英林.数学形态学在图像处理中的应用进展[J].控制理论与应用,2001,18(4):478-482.DAI Qingyun,YU Yinglin. The advances of mathematical morphology in image processing[J]. Control theory and applications,2001,18(4):478-482.
[3]解欢庆.改进的蚁群算法在图像边缘检测中的应用研究[D].兰州:兰州大学,2011.XIE Huanqing. Application research of improved ant colony algorithm in image edge detection[D]. Lanzhou:Lanzhou University,2011.
[4]雒涛,郑喜凤,丁铁夫.改进的自适应阈值Canny边缘检测[J].光电工程,2009,36(11):106-111.LUO Tao,ZHENG Xifeng,DING Tiefu. Improved self-adaptive threshold Canny edge detection[J]. Opto-electronic engineering,2009,36(11):106-111.
[5]张心言,赵冉阳.基于Canny的自适应边缘检测算法及性能评估[J].计算机技术与发展,2015,25(11):32-37.ZHANG Xinyan,ZHAO Ranyang. An adaptive edge-detection algorithm based on Canny and its performance evaluation[J].Computer technology and development,2015,25(11):32-37.
[6]磨少清.边缘检测及其评价方法的研究[D].天津:天津大学,2011.MO Shaoqing. Study on edge detection and its evaluation method[D]. Tianjin:Tianjin University,2011.
[7] BAO Paul,ZHANG Lei,WU Xiaolin. Canny edge detection enhancement by scale multiplication[J]. IEEE transactions on pattern analysis and machine intelligence,2005,27(9):1485-1490.
[8]梁光明,孙即祥,马琦,等.Otsu算法在Canny算子中的应用[J].国防科技大学学报,2003,25(5):36-39.LIANG Guangming,SUN Jixiang,MA Qi,et al. Application of Otsu algorithm in Canny operator[J]. Journal of University of Defense Technology,2003,25(5):36-39.
[9]金刚.自适应Canny算法研究及其在图像边缘检测中的应用[D].杭州:浙江大学,2011.JIN Gang. Research on adaptive Canny algorithm and its application in image edge detection[D]. Hangzhou:Zhejiang University,2011.
[10]刘超,周激流,何坤.基于Canny算法的自适应边缘检测方法[J].计算机工程与设计,2010,31(18):4036-4039.LIU Chao,ZHOU Jiliu,HE Kun. Adaptive edge-detection method based on Canny algorithm[J]. Computer engineering and design,2010,31(18):4036-4039.
[11]傅仲良,李勇.基于梯度幅度直方图和类内方差的边缘提取方法[J].武汉大学学报(信息科学版),2005,30(12):1056-1058.FU Zhongliang,LI Yong. Edge detection based on gradient histogram and variance within clusters[J]. Geomatics and Information Science of Wuhan University,2005,30(12):1056-1058.
[12]邱丽君.Otus图像分割方法的研究与应用[D].济南:山东师范大学,2011.QIU Lijun. Study and application of Otus image segmentation method[D]. Jinan:Shandong Normal University,2011.
[13]李惠光,姚磊,石磊.改进的Otsu理论在图像阈值选取中的应用[J].计算机仿真,2007,24(4):216-220.LI Huiguang,YAO Lei,SHI Lei. Automatic selection of image threshold based on improved Otsu[J]. Computer simulation,2007,24(4):216-220.
[14] MEER P,GEORGESCU B. Edge detection with embedded confidence[J]. IEEE transactions on pattern analysis and machine intelligence,2001,23(12):1351-1365.
[15] RACHID D. Using Canny′s criteria to deriver a recursively implemented optimal edge detector[J]. International journal of computer vision,1987,1(2):167-187.