Using computer vision-enriched discrete choice models to assess the visual impact of transport infrastructure renewal projects: A case study of the Delft railway zone

Abstract

a computer vision-enriched discrete choice model to investigate the impact of the redevelopment of the Delft railway zone on the visual environment. Using computer vision-enriched discrete choice models, we evaluate the changes in the utility levels derived from the visual environment by analysing over 70k street-view images from periods before and after the redevelopment of the railway zone in Delft. We find evidence that the visual appearance of the railway zone has considerably improved after the redevelopment project. This finding highlights the potential of using computer vision-enriched discrete choice models to quantitatively evaluate and monitor changes to the visual environment arising from new transport infrastructure projects.