3D reconstruction based on a fast area-based stereo matching algorithm
被引:0
|
作者:
Gao, Hongwei
论文数: 0引用数: 0
h-index: 0
机构:
Chinese Acad Sci, Robot Lab, Shenyang Inst Automat, Shenyang 110016, Peoples R China
Grad Sch Chinese Acad Sci, Beijing 100039, Peoples R ChinaChinese Acad Sci, Robot Lab, Shenyang Inst Automat, Shenyang 110016, Peoples R China
Gao, Hongwei
[1
,2
]
Wu, Chengdong
论文数: 0引用数: 0
h-index: 0
机构:
Northeast Univ, Sch Informat Sci & Engn, Shenyang 110004, Peoples R ChinaChinese Acad Sci, Robot Lab, Shenyang Inst Automat, Shenyang 110016, Peoples R China
Wu, Chengdong
[3
]
Li, Bin
论文数: 0引用数: 0
h-index: 0
机构:
Chinese Acad Sci, Robot Lab, Shenyang Inst Automat, Shenyang 110016, Peoples R ChinaChinese Acad Sci, Robot Lab, Shenyang Inst Automat, Shenyang 110016, Peoples R China
Li, Bin
[1
]
机构:
[1] Chinese Acad Sci, Robot Lab, Shenyang Inst Automat, Shenyang 110016, Peoples R China
[2] Grad Sch Chinese Acad Sci, Beijing 100039, Peoples R China
[3] Northeast Univ, Sch Informat Sci & Engn, Shenyang 110004, Peoples R China
According to the parallel binocular vision system, a connotative constraint was proved by geometry theory in this paper. That is for any space point projected into two images, along the horizontal direction in image coordinate, there always exists u(1) > u(2). Based on this constraint, a fast stereo cross-correlation algorithm and correlative 3D reconstruction process were introduced. Epipolar rectification, 3D sparse and dense reconstruction results show the validity of the algorithm.