A Practical System for Traffic Sign Recognition

This paper describes a computer vision based system for traffic sign recognition. Two practical aspects, how to achieve an appropriate recognition rate and how to use such a system at night are discussed. The paper will propose to use hierarchical features under a cascade structure for recognition, and to control camera exposure based on multiple detection and recognition results for night. Experimental results with a large amount of videos obtained from actual environments show that the system is competent for further application developments.

Language

  • English

Media Info

  • Media Type: CD-ROM
  • Features: Figures; Photos; References; Tables;
  • Pagination: 8p
  • Monograph Title: ITS in Daily Life

Subject/Index Terms

Filing Info

  • Accession Number: 01150305
  • Record Type: Publication
  • Files: TRIS
  • Created Date: Feb 1 2010 12:34PM