Automated Safety Assessment of Rural Roadways Using Computer Vision

Roadside elements play an important role in the number and severity of crashes. Rigid obstacles (trees, rocks, embankments, etc.), guardrails, clear zones, and side slopes are among the factors that might affect roadside safety. The Federal highway administration (FHWA) presented a rating system to help DOTs and transportation agencies make better decisions about improving road segments. However, the manual process of rating road segments is time-consuming, inconsistent, and labor-intensive. To this end, this project proposed an automated rating system based on images taken from Utah roadways. Utilizing machine learning algorithms and Mandli images, the developed approach employs the FHWA rating system as the primary standard for assessing roadside safety. To provide more detailed information about safety conditions on the roadside, various computer vision algorithms have been developed to detect each roadside feature. The pre-trained models for available clear zone detection and sideslope classification have also been established. A shape-file has been generated by assigning a safety ranking to road segments on five state roads. This product can assist traffic engineers in decision-making to improve road safety by prioritizing projects that address problematic locations. The results show a promising approach to enhancing road safety and preventing crashes.

Language

• English

Project

• Status: Completed
• Funding: $45000
• Sponsor Organizations:

  Utah Department of Transportation

  4501 South 2700 West
  Project Development
  Salt Lake City, UT  United States  84114-8380
• Managing Organizations:

  Utah Department of Transportation

  4501 South 2700 West
  Project Development
  Salt Lake City, UT  United States  84114-8380
• Project Managers:

  Chamberlin, Robert

• Performing Organizations:

  University of Utah, Salt Lake City

  College of Engineering, Department of Civil Engineering
  Salt Lake City, UT  United States  84112-0561
• Principal Investigators:

  Rashidi, Abbas

  Markovic, Nikola

  Mashhadi, Ali

• Start Date: 20220801
• Expected Completion Date: 0
• Actual Completion Date: 20230831

Subject/Index Terms

• TRT Terms: Automation; Computer vision; Highway safety; Image analysis; Ratings; Rural highways; Safety analysis
• Geographic Terms: Utah
• Subject Areas: Data and Information Technology; Highways; Safety and Human Factors;

Filing Info

• Accession Number: 01900454
• Record Type: Research project
• Source Agency: Utah Department of Transportation
• Files: RIP, STATEDOT
• Created Date: Nov 27 2023 12:07PM