摘要
针对基于点特征的单目视觉位姿测量,研究了一种基于不共面5个编码特征点的单目视觉解算算法,利用大靶面高分辨率相机采集运动物体姿态图像,利用目标物体上编码特征点间的几何约束关系,通过正交迭代解算算法完成目标物体的位姿测量。实验结果表明:位移测量不确定度优于0.5mm,姿态测量不确定度优于0.05°,可满足航空航天领域试验测量要求。
Aiming at the pose measurement of monocular vision based on the point characteristics,the algorithm which uses non-coplanar five coded targets is studied.The system uses the high resolution camera with large target surface which can record the pose images of a moving object,then,using the geometric constraint relations of the encoded targets in target object and the orthogonal iteration algorithm,it can realize the pose measurement of the object.The experiment result shows that the displacement uncertainty of measurement is less than 0.5mm and the pose angle uncertainty of measurement is less than 0.05°,and it can meet the measuring demands in aeronautic and aerospace engineering.
作者
袁媛
刘柯
孙增玉
高越
刘华
YUAN Yuan;LIU Ke;SUN Zeng-yu;GAO Yue;LIU Hua(Beijing Aerospace Institute for Metrology and Measurement Technology,Beijing 100076,China)
出处
《宇航计测技术》
CSCD
2019年第6期1-6,共6页
Journal of Astronautic Metrology and Measurement
关键词
单目视觉
物体位姿
正交迭代
编码特征点
算法
Monocular vision
Object position
Orthogonal iteration
Coded target
Algorithm