Fixed Viewpoint Mirror Surface Reconstruction Under an Uncalibrated Camera

This paper addresses the problem of mirror surface reconstruction, and proposes a solution based on observing reflections moving reference plane surface. Unlike previous approaches which require tedious calibration, our method can recover camera intrinsics, poses plane, as well from observed under at least three unknown distinct poses. We first show that 3D be estimated reflection correspondences established between images plane. then form bunch lines correspondences, derive an analytical to line projection matrix. transform matrix its equivalent matrix, propose cross-ratio formulation optimize by minimizing reprojection errors. The is reconstructed optimized constraint. Experimental results both synthetic real data are presented, demonstrate feasibility accuracy method.