机构:
Tsinghua Univ, Sch Informat Sci & Technol, Dept Automat, Beijing 100086, Peoples R ChinaTsinghua Univ, Sch Informat Sci & Technol, Dept Automat, Beijing 100086, Peoples R China
Zhou, Haoyin
[1
]
Zhang, Tao
论文数: 0引用数: 0
h-index: 0
机构:
Tsinghua Univ, Sch Informat Sci & Technol, Dept Automat, Beijing 100086, Peoples R ChinaTsinghua Univ, Sch Informat Sci & Technol, Dept Automat, Beijing 100086, Peoples R China
Zhang, Tao
[1
]
机构:
[1] Tsinghua Univ, Sch Informat Sci & Technol, Dept Automat, Beijing 100086, Peoples R China
来源:
2014 11TH WORLD CONGRESS ON INTELLIGENT CONTROL AND AUTOMATION (WCICA)
|
2014年
关键词:
two-view SfM;
iterative method;
LOW-RANK MATRIX;
D O I:
暂无
中图分类号:
TP [自动化技术、计算机技术];
学科分类号:
0812 ;
摘要:
Structure from motion (SfM) is an important task in computer vision. Two-view SfM, which means SfM from a pair of images, is the foundation of incremental SfM methods and is quite useful in some real applications. Based on the analysis of camera geometry, this paper proposes a two-view SfM algorithm. The proposed algorithm is able to use the rotation information naturally. Simulation and experimental results show that the proposed two-view algorithm is feasible.