Dataset Hub

This dataset includes videos and images collected from YouTube and web crawlers, comprising fire and non-fire data. The non-fire video dataset includes chimney smoke, sunsets, and clouds. The fire video dataset includes flames, smoke, and fires. The fire image dataset includes fire and non-fire images.

This dataset was created by the team for the 2018 NASA Space Apps Challenge, aiming to develop a model capable of identifying images containing fires. The dataset includes two classes: fire images (755 outdoor fire images, some with dense smoke) and non‑fire images (244 natural images such as forests, trees, grasslands, rivers, etc.). The dataset exhibits class imbalance; it is recommended to keep equal numbers of each class in the validation set.

Multi‑Scene Fire and Smoke Detection Benchmark (MS‑FSDB) is a comprehensive and fine‑grained fire and smoke detection benchmark created by Zhejiang University. The dataset contains 12,586 images depicting 2,731 scenes, comprising 3,603 positive samples and 8,983 negative samples. The creation process involved systematic collection of diverse resources from public sources, followed by scene expansion and re‑annotation to ensure data accuracy and consistency. MS‑FSDB includes not only flame detection but also smoke detection tasks, applicable to various complex indoor and outdoor scenarios, and aims to provide strong support for breakthroughs and development in fire detection technology.