The Comparison of Calibration Method of Binocular Stereo Vision System

DOI

10.2991/ic3me-15.2015.171How to use a DOI?

Keywords

Stereo vision, Calibration method, Genetic algorithm, BP neural network

Abstract

Stereo vision is a non-contact method to realize 3D reconstruction of free-form surface, and has a wide range of applications in the field of reverse engineering and virtual reality. The technology of camera calibration has always been the research hotspot and difficulty. In this paper, firstly the experimental system of binocular stereo vision is constructed, and then the stereo vision system is calibrated respectively by using genetic algorithm and BP neural network. In the method of genetic algorithm, the encoding rule of adaptive adjustment of parameter search interval is proposed, which effectively realize the camera calibration of the high dimension and non-linear system of binocular stereo vision. In the calibration method of BP neural network, with the help of CNC precision mobile workstation to obtain the density calibration sample data, by means of BP neural network which simulating the mapping relation between the 3D space and 2D image plane of stereo vision system, the implicit calibration model of binocular stereo vision system is constructed, which avoids the system error because of imperfect mathematical models. The two methods have respective advantages, and are suitable for different application situation.

Copyright

© 2015, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Zhang Ke
AU  - Gao Zhao
PY  - 2015/08
DA  - 2015/08
TI  - The Comparison of Calibration Method of Binocular Stereo Vision System
BT  - Proceedings of the 3rd International Conference on Material, Mechanical and Manufacturing Engineering
PB  - Atlantis Press
SP  - 896
EP  - 901
SN  - 2352-5401
UR  - https://doi.org/10.2991/ic3me-15.2015.171
DO  - 10.2991/ic3me-15.2015.171
ID  - Ke2015/08
ER  -