摘要
针对已有的图像边缘检测算法对大尺度弱边缘的检测效果不理想,提出了一种通过对原图再采样,从中抽取多幅低分辨率子图像,并分别进行边缘检测,再综合各子图像边缘确定最终边缘的新边缘检测算法。由于低分辨率采样下图像高频特征失真较大,而低频特征失真较小的原因,这一算法能很好地抑制图像噪声,有效地提取出图像中大尺度弱边缘。此算法已成功地应用于机器视觉砂轮直径检测系统中。
As it is difficult to detect large-scale weak edge by using classical edge detection algorithm, a new edge detection method was established by extracting some sub-images, detecting edge on each sub-image and integrating the edges of sub-images and the edge of original image to search the edges. Berause of the fact that when image resolution is reduced, the image high-frequency characteristics are distorted and the image low-frequency characteristics are held, this method has noiseproof feature and large-scale weak edge detection property. The method has been used in a grinding wheel computer vision measurement system successfully.
出处
《太原理工大学学报》
CAS
北大核心
2009年第3期291-293,306,共4页
Journal of Taiyuan University of Technology
