Automated Traffic Sign Recognition Using Computer Vision and Deep Learning

Details

• Creators:
  Wang, Yinhai ; Zhang, Shucheng ; Liu, Chenxi ; Jantarathaneewat, Nutvara
• Corporate Creators:
  University of Washington. Washington State Transportation Center (TRAC)
• Corporate Contributors:
  Washington (State). Department of Transportation
• Subject/TRT Terms:
  • Traffic signal control systems
  • Asset management
  • Inventory
  • Deep learning
  • Maintenance
  • Automation
  • Computer vision
  • Video cameras
  • Streets
  • Datasets
  More +
• Publication/ Report Number:
  WA-RD 946.1
• Resource Type:
  Tech Report
• Geographical Coverage:
  United States ; Washington
• Edition:
  Final Report
• Corporate Publisher:
  Washington (State). Department of Transportation
• Description:
  An up-to-date traffic sign inventory is essential for roadway safety, infrastructure maintenance, and support of intelligent road agents, yet traditional manual inspections are inefficient and labor-intensive. This research presents an automated system that leverages deep learning and computer vision to detect, classify, and generate a comprehensive traffic sign inventory from log videos and open street images. To enhance model performance, a new traffic sign dataset tailored to Washington state traffic scenarios was developed and used for training. Additionally, traffic sign condition assessment was explored using six retro-reflectivity-related features. The integrated system consists of four modules-detection, tracking, assessment, and inventory building-each designed to perform a specific function, with the final inventory visualized geospatially. For this system, sparse R-CNN was employed for detection and classification, while a DNN classifier was used for assessment. Experimental results demonstrated high detection accuracy (93.50 mAP@50) and reliable condition assessment (93.56 percent accuracy) while maintaining computational efficiency (25.4 FPS). A comparison with the Washington State Department of Transportation (WSDOT) inventory revealed key discrepancies. By automating traffic sign management, this system provides a scalable, cost-effective solution for transportation agencies to maintain accurate inventories, prioritize maintenance, and enhance roadway safety, thereby ensuring more reliable roadway and seamless mobility for all road users.
• Format:
  PDF
• Funding:
  T1461, Task AA
• Download URL:
  https://rosap.ntl.bts.gov/view/dot/94020/dot_94020_DS1.pdf
• File Type:
  [PDF - 5.53 MB]
• Collection(s):
  US Transportation Collection
• Main Document Checksum:
  urn:sha-512:1d98228454ef5688c3dd15ae11ff0b5cf1c85d81ee4ee0784a4d641d656234df73cb7740a67a1aa33bdbf867553825c73dfb32897106b440d06104d89a34046d