Pedestrians and bicyclists are highly influenced by their perceptions of street environments, making it important to assess the spatial quality of streets and its impact on walkability and bikeability. Current tools for this assessment are limited because they rely on subjective audits or basic metrics. Despite widespread recognition that visual street environments shape walking and biking behavior, no existing method combines automated feature extraction with human perception to objectively quantify street quality.
This project proposes the Visual Street Index for Active Mobility (VISIAM)—an AI-driven framework that integrates computer vision, self-supervised deep learning, and human-perception data to systematically assess bikeability and walkability at the street segment level. Google Street View images from Davis, California and Baltimore, Maryland are analyzed using advanced computer vision models to extract and classify streetscape features. Human-perception evaluations from diverse stakeholder groups then rate street imagery across dimensions including perceived safety, comfort, aesthetics, and traffic stress. The AI-derived classifications and human ratings are integrated to produce a composite bikeability and walkability score for each street segment.
The final output is a citywide, street-level index visualized through interactive maps and dashboards, enabling policymakers to identify gaps, prioritize investments, and explore how specific infrastructure improvements could shift streets from lower to higher quality categories. Key components of the tool will be open-sourced to encourage replication and innovation by researchers, civic tech groups, and agencies.
Universities Involved
Morgan State University
University of Maryland, College Park
Principal Investigators
Hossain Mohiuddin
Celeste Chavis
Vanessa Frias-Martinez
Expected Research Outcomes & Impacts
The project will provide agencies with a cost-effective, replicable alternative to labor-intensive street audits, enabling scenario analysis to guide infrastructure upgrades and support strategic planning, corridor prioritization, and active mobility investments. The VISIAM tool can suggest how low-performing streets might be improved and support agencies’ compliance with federal goals for measurable infrastructure improvement.
The project will involve two undergraduate students, one MS student, and one PhD student, providing hands-on training in artificial intelligence, urban analytics, and transportation planning. The interdisciplinary team comprises faculty from engineering, information science, and transportation technology and policy.
Subject Areas
Active Mobility, Computer Vision, Artificial Intelligence, Urban Analytics, Pedestrian and Bicycle Infrastructure