Mapping street lighting: Illuminance characterization in urban squares in the Brussels Capital Region

Recent literature has acknowledged the critical role street lighting plays in the urban environment, influencing economic growth, human health, biodiversity, pedestrian safety, light pollution, and mobility, amongst others. Street lighting has also become a subject of ongoing discourse due to its energy use and the associated environmental impacts. Balancing the benefits of urban street lighting for environmental wellbeing, while considering energy use, requires identifying appropriate lighting conditions for pedestrians to inform urban policies and implementations, as well as the necessary data to accurately support such decisions. However, most cities fail to document current street lighting conditions through methods such as geospatial, or nighttime aerial imagery analysis, leaving lighting conditions unknown for most municipalities and thereby limiting the potential for their effective amendment, implementation, or assessment. This paper presents the AI LUX tool, which is based on a novel methodology combining street level 360° images, and machine learning computer visions algorithms, to document street level illuminance. The paper also explains the findings of a field campaign conducted in 10 urban squares located in the Brussels Capital Region, where, using illuminance sensors, a geospatial mapping of the street illuminance conditions was performed. This data was then used to calibrate and validate the AI LUX tool. The results show a robust agreement between the field measurements and the illuminance ratio obtained from the AI LUX tool. Given the increasing availability of Google Street View (GSV) and 360º nighttime imagery, the AI LUX tool proposes a scalable workflow to efficiently compute street illuminance.