A Dataset for Visual Classification of Battery Fire and Smoke

Battery fires pose a significant risk across a wide range of applications, including electric vehicles, consumer electronics, and grid-scale energy storage systems. Early detection of fire and smoke is critical to preventing catastrophic failures and ensuring human safety. In this study, we developed a synthetic dataset of battery fire and smoke images in the context of a simple battery pack. The primary application of this dataset is to support the development of a machine learning–based visual classification system capable of accurately detecting battery fires and smoke in real time at an early stage. The intended outcome is a deployable classification system that enhances battery safety through rapid visual identification of hazardous conditions.