摘要
针对传统Canny边缘检测算法自动化程度不高的缺点,提出一种自适应算法。改进的Canny算法通过尺度自适应调整的高斯滤波器对原始图像进行平滑滤波,保留边缘信息并去除噪声的影响。对经过非极大抑制后的图像梯度直方图采用最大类间方差方法对检测和连接边缘的高低阈值进行了自适应设定。理论和实验结果表明,改进的Canny算法简单易实现,满足了原算法的最优化准则,并且提供了边缘的自动化检测能力,得到了良好的检测结果。
The traditional Canny edge detection algorithm is lack of the automation process capabilities in the edge detection. On the basis of the traditional Canny edge detection algorithm, a new method of adaptive Canny algorithm is proposed. The new method Smooth the original image with scale-adaptive Gaussian filter, which retain the edge information and remove the impact of noise. After the Otsu＇ s method applied to image gradient histogram which generated by non-maximal suppression, the new method obtain high and low thresholds automatically in the process of edge detection and connection. With theoretical and experimental results, it is demonstrated that the improved Canny algorithm is simple and easy to realize, fulfilling the optimization criteria of the original algorithm. Moreover, compared with the traditional algorithm, the new method provids the automation process capabilities and show good detection results.
出处
《计算机工程与设计》
CSCD
北大核心
2010年第18期4036-4039,共4页
Computer Engineering and Design
