摘要
主要讨论平移视点单应矩阵的特征理论 ,并应用于 :( 1 )平面规范法向量与平移向量的内积计算 ;( 2 )标准单应矩阵的确定 ;( 3)从未标定图像判定空间平面是否平行 ;( 4 )摄像机自标定 ;( 5)从标定图像 ,计算空间平面间的夹角。本文的结果在机器人视觉中具有重要应用价值。
In this paper, we mainly discuss the eigen-system (eigenvalue and eigensubspace) of homography between two translation viewpoints. And we apply the eigen-system to the following aspects: (1) computing inner-product of normal vector of plane and translation vector; (2) determining standard homography; (3) deciding whether two space planes are parellel from uncalibrated image; (4) self-calibration of camera; (5) computing the angle between two space planes from calibrated image. The results in this paper has important application value in the machine vision.
出处
《安徽大学学报（自然科学版）》
CAS
2000年第3期40-45,共6页
Journal of Anhui University(Natural Science Edition)
基金
国家自然科学基金资助项目!( 698750 0 1 )
安徽省自然科学基金资助项目!( 9833542 5)
关键词
单应矩阵
特征子空间
机器人视觉
平移视点
homograghy
eigenvalue and eigensubspace
angle between two planes
camera self-calibration
