文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > 工程科技 > 信息与通信 > A Camera Calibration Technique Based on OpenCV
侵权投诉

A Camera Calibration Technique Based on OpenCV

介绍了在OPENCV中对摄像机内外参数的标定方法

A Camera Calibration Technique Based on OpenCV

A Camera Calibration Technique Based on OpenCV

Y. M. Wang, Y. Li, and J. B. Zheng

College of Infonnatics & Electronics

Zhejiang Sci-Tech University

Hangzhou, China

Email: ywang@zstu.edu.cn

Zhang puts forward a calibration method [6] based on two­dimensional template, Zhang's calibration method only requires shooting the same calibration template for more than two images from different angles, and it can obtain the internal and external camera parameters, this method does not need to know the specific position and displacement infonnation of planar template movement, and the planar template is easy to make, so this method is simpler and easier than other methods [7]. OpenCV (Intel® open source computer vision library) is open computer vision library which is developed by Intel®, it consists of a series of C functions and a small amount of C++ classes, and realizes many general algorithm about image processing and computer vision, and has a powerful ability to image and matrix operations. This library uses Zhang's calibration method in camera calibration. The camera calibration module of OpenCV provides a good interface for user, and supports for Windows, Linux platfonn, improving the development efficiency, the execution speed of the module is fast, and it has good cross-platfonn portability and can be applied to engineering.

II.

A.

CALIBRATION PRINCIPLE

Abstract-Camera calibration is very improtant in 3D computer vision systems. In this paper, the camera model in OpenCV (open source computer vision library) is discussed, the non-linear distortion of the tangential and radial distortion aberration are considered. Based on these, we introduce the corner extraction of the Camera Calibration Toolbox for Matlab, and an arithmetic of camera calibration based on OpenCV is given. The proposed algorithm can achieve accurate results. It has features such as computational efficiency and good cross-platform portability. It is able to meet the needs of 3D computer vision systems. Experimental results show the feasibility of the proposed method.

distortion

Keywords-camera calibration; OpenCV; computer vision; lens

I.

INTRODUCTION

The design goal of three-dimensional computer VISIon system is to obtain image infonnation from the camera, and calculate the location, shape infonnation of three-dimensional objects, and thus to identify objects in the space. Each point's brightness of image represents the intensity of reflected light of a space object's surface point, and the geometric location of this point in the image correlates with the geometric location of space object's surface corresponding point. The geometric model for camera imaging detennines the relationship between these locations. The geometric parameters of the model are known as the camera parameters, these parameters must be detennined by experiment and calculation, the process of experiment and calculation is known as camera calibration [1]. Camera calibration is very important in computer vision research. An integrated computer vision system can generally be classified into 6 parts [2], including image acquisition, camera calibration, feature extraction, stereo matching, depth detennination, and interpolation. Accurate internal and external camera calibration parameters can not only directly improve the measurement accuracy, but also lay a good foundation for the subsequent stereo image matching and three-dimensional reconstruction [3, 4].

By now, there are many camera calibration methods, such as calibration method based on the three-dimensional calibration reference proposed by Tsai[5], in which calibration of reference object is generally made up of the two planar template which are perpendicular each other, three-dimensional space coordinate of the reference object surface's calibration point need to be known before calibration, so this method needs a expensive accuracy calibration system. Zhengyou

This project is supported by the National Natural Science Foundation of china under grant 60873020, and supported by Zhejiang provincial Natural Science Foundation of china under grants Z1080702

Camera model

The choice of camera model directly influences the fmal calibration results, so it is very important to select the appropriate camera model and make sure internal and external parameters: Internal parameters give a description of the internal optical and geometry characteristics of camera, such as the image center, focal length, lens distortion, etc; External parameters are three-dimensional position and orientation of camera coordinates system compared with the world coordinates system. Commonly used pinhole mode[I,8] which ignores the lens thickness and distortion, it can not be good to reflect the actual situation. The camera model of calibration algorithm in OpenCV is based on pinhole model, and introduces the radial lens distortion and tangential distortion, it more truly reflects the actual distortion of the lens case compared with pinhole model and Tsai's model that only introduces first order radial distortion.

Suppose Pw (Xw, Yw, Zw) is a point of world coordinates system, Pc (Xc, Yc, Zc) is the coordinates of this point in the camera coordinates system, P(x, y) is the physical coordinates of this point in the image plane coordinates system, P (u, v) is the pixel coordinates in the image plane coordinates system, as shown in Fig. 1

403

第1页

你可能喜欢

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email