Tracking algorithm for zoom camera based on compensation matrix

Wei Liu*, Yongtian Wang, Jing Chen, Junwei Guo

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

3 Citations (Scopus)

Abstract

Camera calibration and zoom tracking play an important role in camera tracking, which is widely used in virtual studio, sports broadcast and other concern fields. However, when the focal length becomes longer, the widely used pin-hole model, which can only work well on short-focus lens, is not accurate enough if used in camera calibration and zoom tracking. Accordingly, in order to solve the problem of telephoto lens calibration, a new method was presented by analyzing the intrinsic structure of zoom lens and the relationship between the movement of lens group and the bias of camera pose. A compensation matrix was utilized to compensate the bias of extrinsic parameters of camera when the focal length was changed. The method improved the accuracy of zoom tracking by compensating the bias of camera pose, without changing the camera calibration itself. The experimental results show that the pixel errors are less than 3 pixels when the focal length reaches 80 mm.

Original languageEnglish
Pages (from-to)159-163+168
JournalHongwai yu Jiguang Gongcheng/Infrared and Laser Engineering
Volume40
Issue number1
Publication statusPublished - Jan 2011

Keywords

  • Camera calibration
  • Virtual studio
  • Zoom tracking

Fingerprint

Dive into the research topics of 'Tracking algorithm for zoom camera based on compensation matrix'. Together they form a unique fingerprint.

Cite this