A 3D Reconstruction Algorithm Using Zoomed Images

摘要

A method for shape recovery from zoomed images taken by a stationary camera is proposed in this paper. The imaging model with variant focal lengths is analyzed, and a nonlinear optimization algorithm is applied to solve the focal length parameters and the shape up to a scale factor. To work out this factor, an additional Lambertian reflectance constraint is imposed. Simulated tests have proved the preliminary feasibility of this 3D reconstruction algorithm.