At the moment, urban mobility research and governmental initiatives are mostly focused on motor-related issues, e.g. the problems of congestion and pollution. And yet, we can not disregard the most vulnerable elements in the urban landscape: pedestrians, exposed to higher risks than other road users. Indeed, safe, accessible, and sustainable transport systems in cities are a core target of the UN's 2030 Agenda. Thus, there is an opportunity to apply advanced computational tools to the problem of traffic safety, in regards especially to pedestrians, who have been often overlooked in the past. This paper combines public data sources, large-scale street imagery and computer vision techniques to approach pedestrian and vehicle safety with an automated, relatively simple, and universally-applicable data-processing scheme. The steps involved in this pipeline include the adaptation and training of a Residual Convolutional Neural Network to determine a hazard index for each given urban scene, as well as an interpretability analysis based on image segmentation and class activation mapping on those same images. Combined, the outcome of this computational approach is a fine-grained map of hazard levels across a city, and an heuristic to identify interventions that might simultaneously improve pedestrian and vehicle safety. The proposed framework should be taken as a complement to the work of urban planners and public authorities.


    Access

    Download


    Export, share and cite



    Title :

    Explainable, automated urban interventions to improve pedestrian and vehicle safety



    Publication date :

    2021


    Type of media :

    Preprint


    Type of material :

    Electronic Resource


    Language :

    English




    Interventions to Promote Pedestrian Safety at Urban Intersections

    S. C. Jolibois / J. H. Guerrier / M. Periakaruppan et al. | NTIS | 1997


    Effective Interventions to Reduce Multiple-Threat Conflicts and Improve Pedestrian Safety

    Morris, Nichole L. / Craig, Curtis M. / Van Houten, Ron | Transportation Research Record | 2020


    Automated and Explainable Artificial Intelligence to Enhance Prediction of Pedestrian Injury Severity

    Antariksa, Gian / Tamakloe, Reuben / Liu, Jinli et al. | IEEE | 2025


    Knowledge-based explainable pedestrian behavior predictor

    Melo, Angie Nataly / Herrera-Quintero, Luis Felipe / Salinas, Carlota et al. | IEEE | 2024


    Does facing traffic improve pedestrian safety?

    Luoma, Juha | Online Contents | 2013