Innovating TOD Operationalization: A Safety-Oriented, Technology-Based Framework Leveraging Street-Level Imagery and Computer Vision

Traditional transit-oriented development (TOD) metrics rely on coarse, static datasets that overlook street-level safety elements such as lighting, visibility, and natural surveillance, limiting planners' ability to assess how the built environment influences transit use and perceived safety. This project closes that gap by developing and operationalizing an artificial intelligence (AI)-based TOD measurement framework that integrates Google Street View (GSV) imagery with machine-learning image segmentation to extract fine-grained, pedestrian-experience-oriented indicators of walkability and safety around transit stations. The methodology is demonstrated for all BART stations in the San Francisco Bay Area. Conventional TOD levels are calculated from established indicators (EPA Smart Location Database, WalkScore, parcel-level land use, and accessibility measures), while a parallel image-based measurement is constructed from GSV imagery within 400–800 meter pedestrian catchment areas using PSPNet semantic segmentation to classify streetscape elements. A SafeTOD typology is then developed to classify stations based on measurable safety-related features including lighting, pedestrian and cyclist presence, visibility, and natural surveillance. Traditional and image-based measures are compared using descriptive statistics and spatial analysis (Moran's I, hotspot analysis), and both are incorporated into mode choice and mode shift models to evaluate relationships among SafeTOD scores, crime, and transit ridership. Expected products include a validated, scalable, station-level image-based TOD and SafeTOD dataset, a documented PSPNet-based segmentation workflow and analytical codebase, a SafeTOD classification framework, and empirical evidence comparing image-based and conventional TOD indicators. Outputs are intended for direct use by the California Department of Transportation (Caltrans), metropolitan planning organizations (MPOs), transit agencies, and local planning departments to prioritize station-area investments and strengthen first- and last-mile safety strategies.

Language

• English

Project

• Status: Active
• Funding: $238,753.00
• Contract Numbers:

  69A3552348323

• Sponsor Organizations:

  Office of the Assistant Secretary for Research and Technology

  University Transportation Centers Program
  Department of Transportation
  Washington, DC  United States  20590
• Managing Organizations:

  Howard University

  2400 6th Street, NW
  Washington, DC  United States  20059
• Project Managers:

  Bruner, Britain

• Performing Organizations:

  San Jose State University

  1 Washington Sq
  San Jose, California  United States  95192
• Principal Investigators:

  Kianmehr, Ayda

• Start Date: 20260810
• Expected Completion Date: 20270504
• Actual Completion Date: 0
• USDOT Program: University Transportation Centers

Subject/Index Terms

• TRT Terms: Computer vision; Image analysis; Machine learning; Mode choice; Pedestrian safety; Quantitative metrics; Rail transit stations; Transit oriented development; Walkability
• Geographic Terms: San Francisco Bay Area
• Subject Areas: Data and Information Technology; Pedestrians and Bicyclists; Planning and Forecasting; Public Transportation; Safety and Human Factors; Terminals and Facilities;

Filing Info

• Accession Number: 01999721
• Record Type: Research project
• Source Agency: Research and Education for Promoting Safety (REPS) University Transportation Center
• Contract Numbers: 69A3552348323
• Files: UTC, RIP
• Created Date: Aug 19 2026 11:50AM