摘要
Canny边缘检测算法中有个可以改变的自由参数,该参数具有内在的尺度意义。根据图像自适应地确定该参数所在的尺度集合,在每个尺度上进行Canny边缘检测,然后组合它们的输出结果以获得理想的边缘。自适应多尺度Canny边缘检测算法可以得到抗噪性能和细节保持兼顾的良好边缘。实验结果表明,与Canny边缘检测相比,该算子检测的图像边缘比较清晰,具有较高的信噪比,而且用来检测含白噪声图像的效果也相当理想。
The Canny detector has a changeable free parameter that has the buih-in sense of scale. The filter scale set that the parameter belongs to is determined adaptively according to the original image. The edgestrength maps were obtained using the Canny edge detector in every scale of the set, and the edge strengths of different scales were combined to get the ideal edge. The multi-scale edge detector performs well in beth the noise resisting and the detail maintaining. Comparative study revealed its superiority over the Canny edge detector in definition of image edge signal-to-noise ratio. So it can be applied to detection of noisy images.
出处
《山东建筑大学学报》
2006年第4期360-363,共4页
Journal of Shandong Jianzhu University