Electronic Thesis and Dissertation Repository

Degree

Master of Science

Program

Computer Science

Supervisor

Steven Beauchemin

2nd Supervisor

Michael Bauer

Co-Supervisor

Abstract

The availability of affordable depth sensors in conjunction with common RGB cameras, such as the Microsoft Kinect, can provide robots with a complete and instantaneous representation of the current surrounding environment. However, in the problem of calibrating multiple camera systems, traditional methods bear some drawbacks, such as requiring human intervention. In this thesis, we propose an automatic and reliable calibration framework that can easily estimate the extrinsic parameters of a Kinect sensor network. Our framework includes feature extraction, Random Sample Consensus and camera pose estimation from high accuracy correspondences. We also implement a robustness analysis of position estimation algorithms. The result shows that our system could provide precise data under certain amount noise.

Keywords

Kinect, Multiple Camera Calibration, Feature Points Extraction, Correspondence, RANSAC

Share

COinS