摘要
Canny边缘检测算法是一种综合评价比较好的边缘检测算法,但在应用时由于用高斯滤波器对图像进行平滑,往往引起定位精度下降并使角点变圆。针对上述问题,采用各向异性高斯滤波对Canny算法进行了改进。实验结果表明,改进后的算法在对噪声的鲁棒性、定位精度和单一响应方面均优于原始的Canny算法。
The Canny edge detection algorithm is an edge detector with good performance as a whole. However, image smoothing by Gaussian filtering may lead to higher edge location error and comer roundness in application. We improved the Canny algorithm by replacing the Gaussian convolution with anisotropic Gauss filtering. Experiment results show that the improved Canny algorithm performs better than the original algorithm in robustness, accuracy and single response.
出处
《电光与控制》
北大核心
2006年第6期53-55,共3页
Electronics Optics & Control