摘要
提出一种基于图像序列的空间直线重建新方法,这种方法是在空间点算法的基础上改进的,不需要特征点的参与。算法首先对直线测量矩阵估计射影深度,采用SVD(Singular Value Decomposition)分解矩阵,并最后应用变换,使得分解结果真实有效。在估计射影深度时运用了共轭法迭代,使得射影深度的估算更准确。
A new method about llne rccomtruction of space based on image sequences is proposed in this paper. This method comes from improvement of space point reconstruction and doesn＇t use feature point correspondence. The method estimates projective depth of llne measure matrix, does SVD and applies transform to make the result validate. Conjugate iterative to estimate projective depth is applied in the method. The projective depth will be more exact.
出处
《计算机与现代化》
2006年第12期63-64,68,共3页
Computer and Modernization
关键词
直线重建
SVD
射影深度
图像序列
line reconstruction
SVD
projective depth
image sequences
