3D reconstruction based on monocular image sequences

Abstract

Aiming at the complex operation of active 3D reconstruction technology, which is easily affected by external environment and equipment, this paper adopts a series of monocular image sequences taken by cell phone cameras from different angles to realize 3D scene reconstruction. Firstly, Structure-From-Motion (SFM) algorithm is used for sparse point cloud reconstruction, and the similarities and differences between incremental and global SFM algorithms for sparse reconstruction are compared; secondly, Multi-View-Stereo (MVS) is used for dense point cloud reconstruction, and finally, the surface and texture information of the object is recovered. The surface and texture information of the objects are recovered to achieve the reconstruction of 3D scenes. The experimental results show that the method can achieve better 3D reconstruction of monocular image sequences.