As cities evolved, residential development became the dominant land use, thus playing an important role in the development of the urban form. With low levels of available land in proximity to urban cores, a great majority of housing starts are required to locate further from the city center than those preceding them.
This case study from Lethbridge, Canada aims to link urban sprawl to national housing finance policy. The urban sprawl component of the study was measured using ATLAS, an AI platform developed by SPH#nbsp;Engineering. The platform’s capabilities supported extracting building footprints from aerial imagery to show the change over time.
The City of Lethbridge is the third-largest city in the Province of Alberta, currently home to approximately 101'482 residents according to the 2019 municipal census. Located in the southern part of the province, the city’s corporate boundaries encompass 122 km2.
Aerial imagery was obtained for a 10-year period to investigate the timing and the rate of sprawl. Measuring land cover is a common method of evaluating the level of urban sprawl. Building footprints, required for the urban sprawl analysis, were extracted using SPH Engineering’s ATLAS platform. ATLAS was instrumental in detecting buildings from the georeferenced aerial images provided by the City of Lethbridge, which were subsequently exported as shapefiles for use in ArcGIS Pro and ArcMap for further analysis.
Price Leurebourg, Master of Science candidate, Heriot-Watt University
The academic study of urban planning
Upload of aerial imagery
Map storage and processing on ATLAS
Initial machine-learning detections to train the detector
Running detector on 6 city-wide aerials to detect all buildings
ATLAS allowed the author to store and process the orthomosaics while minimizing the processing time and providing computing power. This led to the possibility of visualizing both the location and intensity of development activity over time.