Montreal Traffic Camera Images with Segmentation and Object Detection Labels

Montreal traffic camera video archives provide annotated images for computer vision tasks. The dataset includes source images with accompanying semantic segmentation masks covering 13 classes and XML files for object detection bounding boxes covering 5 classes. It is published by the Government and Municipalities of Québec under a CC-BY-4.0 license.

Use Cases

• Train semantic segmentation models based on the 13 provided classes, including posts, traffic signs, vehicles, and pedestrians.
• Develop object detection algorithms based on bounding box annotations for vehicles, pedestrians, cyclists, buses, and construction objects.
• Analyze urban infrastructure and traffic flow patterns based on annotated camera imagery.
• Benchmark model performance on a real-world, city-scale traffic monitoring dataset.

Strengths

• Annotations are provided for two distinct computer vision tasks: semantic segmentation and object detection.
• Semantic segmentation includes 13 distinct classes, offering a detailed breakdown of scene elements.
• Object detection annotations cover 5 specific classes relevant to urban traffic monitoring.

Limitations

• The number of images, file size, and specific column-level metadata are unknown, limiting suitability assessment.
• The description metadata is limited; actual data quality and annotation consistency require manual inspection after download.

Provenance

Source

Government and Municipalities of Québec

Collection Method

Images extracted from recordings of road observation cameras.

Freshness

Last updated 2026-04-17 15:47:47.035123; freshness should be verified.

Geography

Montreal, Québec, Canada