Automated Inventory and Analysis of Highway Assets

The goal of this project is to develop a system which can automatically detect and localize the traffic signs along the roads, in order to provide a comprehensive sign database for sign inventory purpose. The project work focuses on systems integration of existing technologies to create reliable applications for high performance road data asset inventory. The fundamental methods used for feature extraction, object tracking and retrieval in these applications are well-established. Exploration of the advanced, state-of-the-art or experimental algorithmic development was also made during the project. The work also addresses the issue of incorporating parallel computing in order to achieve the real-time characteristics of the system. This report describes the research work on system design, data acquisition, image processing, object tracking, feature extraction, and database recording. The system developed in this project is capable of reliably capturing the Right-Of-Way (ROW) images. The software developed in this project can detect and recognize most of the commonly used road signs with a 99% accuracy rate in a driving speed up to 60 mph.

Language

  • English

Media Info

  • Media Type: Digital/other
  • Edition: Technical Final Report
  • Features: Figures; Photos; References; Tables;
  • Pagination: 32p

Subject/Index Terms

Filing Info

  • Accession Number: 01049506
  • Record Type: Publication
  • Report/Paper Numbers: MBTC 2065
  • Files: UTC, NTL, TRIS, USDOT
  • Created Date: May 6 2007 8:20PM