摘要
提出了一种新型的测量图像快速亚像素边缘检测方法。首先通过计算机模拟仿真,找出图像边缘处灰度落差与边缘点偏移之间的规律性,建立了一个由灰度落差得到边缘亚像素定位点的查找表;然后结合改进的Sobel检测算子和十字窗检测法将图像边缘检测精度达到像素级,并应用查找表实现了亚像素级的检测精度。实验表明:该方法不仅得到了较高的检测精度,定位精度为0.3pixels,而且可以大大地提高检测速度。
A novel fast sub-pixel edge detection method for image measurement is proposed. Firstly the author finds out the relation between gray difference and sub-pixel edge locating points through computer simulating, and establishes a lookup table which can gain sub-pixel edge locating points according to gray difference. Secondly the author combines the improved Sobel operator and the crossed window to inspect image edge, which makes its detection precision reaching pixel level, and apply lookup table to achieve sub-pixel level detection precision. The experimental results show that such method not only gains high detection precision (the positioning accuracy amounts to 0. 3 pixels) ,but also advances detection rate by a long way.
出处
《计算机工程与应用》
CSCD
北大核心
2008年第14期59-61,171,共4页
Computer Engineering and Applications
基金
山西省自然科学基金(the Natural Science Foundation of Shanxi Province of China under Grant No.20051043)
中北大学科学基金资助项目(the Science Foundation of North University of China under Grant)
关键词
亚像素边缘检测
查找表
计算机视觉
sub-pixel edge detection
lookup table
computer vision