摘要
为了恢复图像遮挡点,假设相机为正投影模型,提出了一种基于1维子空间的遮挡点恢复方法,该方法利用第1幅图像点构成的2个行向量外加1个行向量可以组成三维空间点构生成的子空间的特性,线性迭代地求取遮挡点的真实图像位置。此方法最大的优点是在整个求解过程中都是线性的。收敛性实验表明了该方法具有很好的收敛性能,同时,和Eriksson方法及Ma方法比较显示了本文方法具有较好的收敛精度。
To recover occlusions, an occlusion recovery method based on 1D subspace is presented using orthographic projection in the paper. The method relies on the facts that the rows composed of all the image points span the same linear subspace as the rows composed of the 3D space points and that the basis of the subspace can consist of the two row vectors composed of the points of the first image and a row vector that is orthogonal to the former. The occlusions are obtained by linear iteration. The important feature is that the method is linear. Simulation experiment results show that the method is robustness and has good convergence. At the same time, the comparisons with Eriksson＇ s method and Ma＇s method reveal that the method has high convergence precision.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2011年第9期2029-2033,共5页
Chinese Journal of Scientific Instrument
基金
国家自然科学基金(No.60805016)
高等学校博士学科点专项科研基金(新教师基金课题)(No.200807181007)资助项目
关键词
遮挡点恢复
正投影
一维子空间
occlusion recovery
orthographic projection
1 D subspace
