摘要
本文介绍一种对图像中已知形状的目标进行亚像素量级边缘检测的新方法.与其它基于边缘法向信息的亚像素边缘检测方法不同,本方法利用了边缘切向的信息.文中介绍了方法的原理,并研究了目标尺寸、目标采样位置和目标朝向对该方法性能的影响.这一新方法已实际应用并与其它方法用真实图进行了比较,结果表明本方法在实用中能提供更为精确的亚像素边缘.
A new method for locating edges of known-shape objects in digital images to subpixel accuracy is presented. In contrast to other subpixel edge detection methods which are based on the information along the normal direction of edges, this method uses the information along the tangent direction of edges. The principles of the method are described. The effects of object size, object sampling location and object orientation on the performance of this method are studied. This method has also been applied to real applications aad compared with other existing methods with real images. The results show that this method can provide more accurate subpixel edge location in real applications.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
1997年第1期69-74,共6页
Pattern Recognition and Artificial Intelligence
基金
国家教育委员会留学回国人员科研资助费(1994660)
博士点基金(199416)
关键词
机器视器
图像处理
边缘检测
亚像素
自动检测
Machine Vision, Image Processing, Edge Detection, Subpixel, Perfonnance Analysis.
