Critical Assessment of Automatic Traffic Sign Detection Using Three-Dimensional LiDAR Point Cloud Data

Traffic signs are important roadside appurtenances that provide critical guidance to road users including regulations, destinations, and safety related information. Traffic signs need to be inventoried by transportation agencies for asset management and maintenance purposes. However, the traditional method of manually inventorying signs is dangerous, labor-intensive, and time-consuming. There is a need for a safer and more effective traffic sign inventory method. This paper is the first that critically assesses the use of an automatic method for traffic sign detection using 3D LiDAR point cloud data in support of traffic sign inventory. The contribution of this paper is three-fold: 1) it presents an automatic method for traffic sign inventory using 3D LiDAR point cloud data; 2) it critically assesses the performance of the presented method in terms of detection rate and false negative/false positive cases; 3) it suggests the adequate parameter values to achieve a good traffic sign detection rate. Actual data, collected on Interstate 95 (major arterial) and 37th Street in Savannah, Georgia (local road), is used to assess the performance. Results show that the presented method can correctly detect 94.0% and 91.4% of the traffic signs on both roadways, respectively, with less than 7 false positive cases. The results demonstrate that the presented method using 3D LiDAR point cloud data is promising for traffic sign inventory. Future research directions are recommended.

• Supplemental Notes:
  • This paper was sponsored by TRB committee AHD55(1) AHD55 Paper Review
• Corporate Authors:

  Transportation Research Board

  500 Fifth Street, NW
  Washington, DC  United States  20001
• Authors:
  • Ai, Chengbo
  • Tsai, Yichang (James)
• Conference:
  • Transportation Research Board 91st Annual Meeting
  • Location: Washington DC, United States
  • Date: 2012-1-22 to 2012-1-26
• Date: 2012

Language

• English

Media Info

• Media Type: Digital/other
• Features: Figures; Photos; References; Tables;
• Pagination: 15p
• Monograph Title: TRB 91st Annual Meeting Compendium of Papers DVD

Subject/Index Terms

• TRT Terms: Cloud computing; Data collection; Detection and identification system applications; Inventory; Laser radar; Traffic signs
• Subject Areas: Data and Information Technology; Highways; I70: Traffic and Transport;

Filing Info

• Accession Number: 01373720
• Record Type: Publication
• Report/Paper Numbers: 12-3214
• Files: TRIS, TRB
• Created Date: Jun 25 2012 2:45PM