摘要
本文提出一种由模型表面与特征点的拟合估计三维物体位姿参数的方法。这种方法用广义距离代替精确距离进行，或到曲面距离的计算，用修正的拟牛顿法对位姿参数进行迭代估计。计算机模拟结果验证了这一方法的有效性、准确性和稳定性。
This paper presents a method of 3D pose estimation by fitting model surfaces to fea-ture points.In this method instead of the exact distance a generalized distance is used to calculate the distances from the surfaces of the object to the feature points and a revised version of quasi-Newton methods is used to solve for the pose parameters iteratively,Computer simulation re-sults are given to show the efficiency,accuracy,and robustness of the method.
出处
《系统工程与电子技术》
EI
CSCD
1994年第9期40-46,共7页
Systems Engineering and Electronics
关键词
计算机视觉
参数估计
拟合估计
Computer vision,Parameter estimation,Feature selection,Modol.
