Research on 3D Reconstruction Based on Multiple Views

Chen Xu, Qingfeng Wu, Shengzhe Wang

2018 · 17 citations · 6 references

Concepts

Abstract

The acquisition of 3D models is a fundamental research in the field of computer graphics and computer vision. However, it is very complicated and expensive to build 3D models manually by using modeling software, such as 3DMAX and Maya. Therefore, how to get 3D models directly and quickly from the real world is a hot research topic in this field. Multi-view based 3D reconstruction techniques include feature point extraction and matching, camera calibration, sparse point cloud reconstruction, dense point cloud reconstruction, Poisson surface reconstruction, texture mapping and many other techniques. This paper will focus on the main algorithms of 3D reconstruction algorithms - Structure from Motion, Multi-View Stereo and Poisson Surface Reconstruction. Based on the above research, this paper designed a 3D reconstruction system including the above algorithms. The system takes a variety of views of an object as input, and finally outputs the 3D model of the object. Finally, this paper verified the effectiveness and feasibility of the system through experimental analysis.

References

6