Quebec Building Footprint Vector Polygons with AI Extraction

A continuous vector polygon layer of building footprints in Quebec, created from a bird's eye view. The dataset integrates geometries from various partnerships, produced using artificial intelligence and automated extraction algorithms, and includes raw (NC-0) and manually validated (NC-1) data.

Use Cases

• Train geospatial AI models on raw vector geometries (NC-0) to automate building footprint extraction from aerial imagery.
• Validate and refine AI-extracted building footprints against manually validated geometries (NC-1) for quality assessment.
• Analyze the built environment by integrating building footprint polygons with other municipal or provincial geospatial data.
• Assess geometric data quality variations across regions based on the documented levels of completeness (NC).

Strengths

• Data includes both raw (NC-0) and manually validated (NC-1) building geometries for quality comparison.
• Geographic coverage corresponds to the entire province of Quebec, representing a significant spatial area.
• Produced through collaborative work involving Laval University, municipal partners, and federal/provincial agencies.

Limitations

• Geometric quality varies by the level of completeness (NC) and may differ from ground reality in representation and authenticity.
• The dataset's completeness and geographic coverage are dependent on the availability of new source data for extension.

Provenance

Source

Center for Research in Geospatial Data and Intelligence at Laval University, municipal sector, Natural Resources Canada, and Quebec ministries and agencies.

Collection Method

Created from a bird's eye view using artificial intelligence and automated extraction algorithms, integrating geometries from various partnerships.

Time Range

null

Freshness

null

Geography

Quebec, Canada