摘要
灰度图像中边缘定位受采样密度限制，亚像元精度度量方法在需要高精度的应用中受到重视。要想知道到底需要多高的分辨率才能达到给定精度指标，就有必要分析特征定位方法的最坏情况的精度。文中用分析和实验的方法给出了基于空间矩的边缘定位方法，对于边缘点、边缘直线以及直边缘区域的最坏情况定位精度分析。分析和讨论的结果不但可以用来确定已知形状位置估计的最大误差，而且还可用于对准的形状几何参数，以获得较高的精度。
The location precision of an edge measured in a greylevel image is limited by the sam- pling resolution . For those cases, subpixel measurements are at a premium in higher precision applica- tion. It' s useful to know how high a resolution is required for a specified spatial precision. The worst- case precision error ought to be determined . The analysis and experiment results of worst -case preci- sion error for a point, an edge, and a straight-edged shape are proposed in this paper. The conclusion can be used not only to quantify worst-case precision error of measurement of location of known shapes, but also to determine the geometric parameters, so that its shape minimizes the worst-case pre- cision error.
出处
《红外与激光工程》
EI
CSCD
1998年第2期6-10,共5页
Infrared and Laser Engineering
