Complete Street Data Collection and Assessment using Machine Vision

Poorly designed streetscapes contribute to traffic crashes, poor economic performance, neighborhood decline, and many other societal impacts. Designing streets that are focused on the users, complete streets, can have a significant impact on safety and quality of life in our communities. A primary challenge of designing and improving streetscapes is understanding where improvements are needed and what really contributes to a complete street. The purpose of this project is to develop a robust machine vision data collection program that can recognize, identify, and understand streetscape assets and define their impact on the complete street environment. Our team already has extensive experience applying machine vision models to various infrastructure problems, including asset identification and rating the safety of roadside hazards. We also have proven success commercializing these and other technologies. Our team proposes to use this SBIR funding to expand our models to capture assets in urban contexts and be able to identify areas with high-quality complete streetscapes and areas where improvements are needed to better serve users. The model will be built to work with a variety of data sources including street-level imagery and aerial imagery.