Waste Dumpsites Drone Imagery for Computer Vision Detection

2,115 drone images of illegal waste dumpsites collected as part of the Raven Scan project. Images have a resolution of 1280 x 1280 pixels and are captured from a nadir perspective. The dataset is hosted by INS-IntelligentNetworkSolutions and was last updated on December 5, 2024.

Use Cases

• Train object detection models to identify illegal waste dumpsites based on drone imagery.
• Benchmark model performance for environmental monitoring tasks using the provided train/valid/test splits.
• Develop automated systems for regulatory enforcement based on the annotated dumpsite locations.
• Analyze the visual characteristics of waste sites from a consistent nadir perspective.

Strengths

• Contains 2,115 annotated drone images.
• Images have a consistent 1280 x 1280 pixel resolution and nadir perspective.
• Annotations are provided in COCO format with separate train, validation, and test splits.

Limitations

• The specific geographic origin and time range of the imagery are unknown.
• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset's scale in terms of total annotations or unique sites is not detailed.

Provenance

Source

INS-IntelligentNetworkSolutions via the Raven Scan project.

Collection Method

Collected via drone imagery.

Freshness

Last updated 2024-12-05 03:15:59; freshness should be verified.