3,771 labeled satellite images from Landsat-8 and GOES-16 sources, split into training, validation, and test subsets. The dataset was created by Aydin Ayanzadeh for early wildfire detection and smoke analysis, with images resized to 416 × 416 pixels. It was last updated on April 20, 2026.
Use Cases
- Training object detection models for wildfire identification based on labeled satellite imagery.
- Analyzing smoke transport patterns based on high-frequency GOES-16 imagery.
- Developing early warning systems for active fires based on multispectral Landsat-8 data.
- Benchmarking AI models for environmental monitoring based on the provided train/validation/test split.
Strengths
- 3,771 images provide a substantial base for model training.
- Data is pre-split into training (70%), validation (15%), and test (15%) subsets.
- Images are sourced from complementary satellite systems (Landsat-8 and GOES-16) offering different spatial and temporal resolutions.
- All images are preprocessed to a uniform resolution of 416 × 416 pixels.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- The geographic scope is implied but not explicitly stated, which may limit suitability assessment for specific regions.
Provenance
- Source
- Landsat-8 and GOES-16 satellite imagery from NASA Earth observation sources.
- Collection Method
- Publicly available satellite data was collected, preprocessed, and annotated into 'wildfire' and 'smoke' object classes.
- Freshness
- Last updated 2026-04-20 20:10:09; freshness should be verified.
- Geography
- Likely focuses on the Americas, given GOES-16 coverage, but not explicitly stated.