摘要
针对传统Canny边缘检测算法对噪声鲁棒性较差的问题,根据Canny算子滤波器的尺度较小可以检测较小的边缘,但对噪声的抑制能力较差,反之检测较大的边缘对噪声的抑制能力较强的特点,提出了一种多尺度下的噪声图像边缘检测方法。运用Canny算子在不同尺度下检测边缘的优点,对含噪图像进行不同尺度的Canny边缘检测,对不同尺度边缘检测结果进行跟踪实现噪声图像的边缘检测。仿真实验结果表明,与传统的方法相比,改进的算法能较清晰地提取出图像的边缘,降低高斯噪声对图像边缘的影响。
Aiming at poor robustness of the traditional canny edge-detection algorithm,based on that the Canny operator filter with small scale could detect smaller edge but have poor restraint capability on noise,otherwise detect larger edge and have stronger restraint capability on noise,a multiscale edge detection algorithm is presented.Applied the edge-detection advantage of canny operators to multiscale, firstly,implements the canny edge-detection of different scale to the noisy images,and then follows it to realize the edge-detection of noisy images by edge track.Finally compared to the traditional method,the simulation results indicated that the multi-scale algorithm can effectively remove the noise and get relatively complete edge information.
出处
《计算机工程与设计》
CSCD
北大核心
2011年第1期228-231,共4页
Computer Engineering and Design
