Robust multi-view L2 triangulation via optimal inlier selection and 3D structure refinement
作者：
Highlights：
• The uncertainty of epipolar transfer is analytically characterized.
• Two new algorithms to calculate the error bounds of 3D structure are proposed.
• A new robust formulation of direct multi-view triangulation is proposed.
• Our method can identify more inliers than existing robust triangulation strategies.
• Our method can reconstruct more unambiguous 3D points with higher accuracy.
摘要
Highlights•The uncertainty of epipolar transfer is analytically characterized.•Two new algorithms to calculate the error bounds of 3D structure are proposed.•A new robust formulation of direct multi-view triangulation is proposed.•Our method can identify more inliers than existing robust triangulation strategies.•Our method can reconstruct more unambiguous 3D points with higher accuracy.
论文关键词：Multi-view triangulation,Optimal inlier selection,3D structure error bounding,Differential evolution
论文评审过程：Received 15 February 2013, Revised 22 January 2014, Accepted 24 March 2014, Available online 2 April 2014.
论文官网地址：https://doi.org/10.1016/j.patcog.2014.03.022
