Explore high-quality datasets for your AI and machine learning projects.
The FireRisk dataset is a remote‑sensing fire‑risk classification collection. It contains RGB three‑band images of size 320 × 320 pixels at 1‑meter resolution. The dataset comprises 91,872 images and 101,878 tree annotations across seven land‑cover classes: high, low, moderate, non‑burnable, very high, very low, and water. The source imagery is NAIP aerial photography.
