北京邮电大学学报

  • EI核心期刊

北京邮电大学学报

• 论文 • 上一篇    下一篇

路况PTZ摄像机自动标定方法

李勃   

  1. 南京大学电子系1132#
  • 收稿日期:2009-04-13 修回日期:1900-01-01 出版日期:2009-04-28 发布日期:2009-04-28
  • 通讯作者: 李勃
  • 基金资助:
     

Automatic Calibration Method for PTZ Camera

    

  1.  
  • Received:2009-04-13 Revised:1900-01-01 Online:2009-04-28 Published:2009-04-28
  • Supported by:
     

摘要: 为解决路况pan/tilt/zoom(PTZ)摄像机内外参数实时多变,绝对距离难以求解的问题,提出了一种基于路况图像的自动化标定方法: 建立全新的路况PTZ摄像机成像模型; 根据图像中提取的路面点的相对距离关系,推导模型中各未知参数的求解方法,实现摄像机参数的“白盒”标定,并给出了自动化标定的实现方案; 摘录实际路况试验结果,计算标定误差. 该算法只需要路面场景图像即可实现自动标定,无须人为干涉和配置,其计算简单,灵活性好,非常符合智能交通监控系统的需求.

关键词: 视频监控, PTZ摄像机, 摄像机标定, 角点检测

Abstract: Considering that the intrinsic and extrinsic parameters of road side pan/tilt/zoom (PTZ) camera are variable, the absolute distance of object to camera is difficult to solve. An automatic calibration method for PTZ camera based on analyzing traffic video is presented. A model of traffic camera is given. By using the relative distances of points on road, a method is established to solving all the unknown parameters in the model. An automatic calibration implementation procedure is given. Experiment shows the calibration errors. During the procedure, only images of the traffic scene are needed without any manual intervention and configuration. This method has advantage of simple calculation and good flexibility, the demand of intelligent transportation surveillance system can be satisfied.

Key words: video surveillance, pan/tilt/zoom camera, camera calibration, corner detection

中图分类号: