A Real-World Benchmark for Early Wildfire Detection Using Sequential Data with the PyroNear Dataset

Early wildfire detection (EWD) is of the utmost importance to enable rapid response efforts and thus minimize the negative impacts of wildfire spreads. To this end, we present PyroNear${}_{2025}$ , a new dataset composed of both images and videos, allowing for the training and evaluation of smoke plume detection models, including sequential models. The data is sourced from the following: (i) web-scraped videos of wildfires from public networks of cameras for wildfire detection in-the-wild, (ii) videos from our in-house network of cameras, and (iii) a small portion of synthetic and real images. This dataset includes around 150,000 manual annotations on 50,000 images, covering 640 wildfires; PyroNear${}_{2025}$ surpasses existing datasets in size and diversity. It includes data from France, Spain, Chile, and the United States. Finally, it is composed of both images and videos, allowing for the training and evaluation of smoke plume detection models, including sequential models. We ran cross-dataset experiments using a lightweight state-of-the-art object detection model, similar to the ones used in real-world applications, and found that the proposed dataset is particularly challenging, with an F1 score of around 70%, but it is more stable than existing datasets. Finally, its use in concordance with other public datasets helps to reach higher results overall. Last but not least, the video part of the dataset enables another technical contribution, as it can be used to train a lightweight sequential model, improving global recall while maintaining precision for earlier detections. The output of this work has real-life implications, as it is used to automatically detect wildfires, with our models running on Raspberry Pi in several countries. We will make both our code and data available online.