摘要

Reconstructing a 3D model of a scene from a set of 2D images is a fundamental problem in computer vision with many applications. A new technique of cameras reconstruction based on unify the world coordinate system and L∞ methods is presented inthis paper. This problem can be solved in three steps: first estimate camera rotations and then translations are estimated using a standard technique based on Second Order Cone Programming. At last, L∞-norm methods are used to realize cameras reconstruction. The method have been implemented in Matlab and the resulting toolbox has been made publicly available. The improved solutions have been validated on real images with excellent performance.

  • 出版日期2009

全文