一种改进的Canny边缘检测自适应算法

An improved adaptive Canny edge detection algorithm

针对传统Canny算子在高斯平滑时,对图像边缘区域的影响,以及需要预先设定高低阈值的缺点,本文作者提出了一种新的Canny边缘检测自适应算法.首先,该算法根据高斯滤波原理和特征统计分析的方法来自适应地选择滤波函数、滤波窗口;然后,对梯度的求解采取增强中间像素影响力的方法来提高边缘检测的精度;最后,结合非最大抑制后的梯度幅值图和阈值分割的差分理念,提出了一种新的自适应算法来选择高低阈值.实验结果表明,该方法能够得到较好的边缘检测效果.

The paper brings forward a new adaptive Canny edge detection algorithm which aimed the image edge areas effect when traditional Canny operator under Gaussian smooth and the disadvantages of threshold level pre-set. At first, the algorithm adaptively select filter function according to Gaussian filtering principle and statistical analysis of characteristics. Then, for gradient solving, it takes the way enhancing influence of intermediate pixel to improve the accuracy of edge detection. Finally, this paper proposes a new adaptive algorithm to select threshold by combining gradient magnitude diagram in nonmaximum suppression and differential idea of threshold segmentation. The experimental results show that the method can get better effect of edge detection.