WildfireX-SLAM: A Large-scale Low-altitude RGB-D Dataset for Wildfire SLAM and Beyond

TL;DR

WildfireX-SLAM tackles the lack of outdoor wildfire SLAM data by introducing a large-scale synthetic dataset built with Unreal Engine 5 and AirSim, featuring low-altitude UAV RGB-D imagery over a 16 square-kilometer forest and multi-modal data (depth, thermal, normals) under controllable wildfire, weather, and lighting. The authors develop a data-construction pipeline that integrates PCG-based forest generation with Niagara-based fire simulation and an enhanced AirSim for automated recording, enabling realistic sim-to-real study opportunities. They benchmark state-of-the-art 3D Gaussian Splatting SLAM methods (MonoGS, SplatAM) on two benchmark subsets and four wildfire-driven challenges, revealing how dynamic fires, smoke, lighting, and fog degrade localization and map quality, while highlighting the relative strengths of different approaches. The work provides a publicly available dataset and framework to support wildfire emergency response, forest mapping, and swarm-based perception research in dynamic, safety-critical environments.

Abstract

3D Gaussian splatting (3DGS) and its subsequent variants have led to remarkable progress in simultaneous localization and mapping (SLAM). While most recent 3DGS-based SLAM works focus on small-scale indoor scenes, developing 3DGS-based SLAM methods for large-scale forest scenes holds great potential for many real-world applications, especially for wildfire emergency response and forest management. However, this line of research is impeded by the absence of a comprehensive and high-quality dataset, and collecting such a dataset over real-world scenes is costly and technically infeasible. To this end, we have built a large-scale, comprehensive, and high-quality synthetic dataset for SLAM in wildfire and forest environments. Leveraging the Unreal Engine 5 Electric Dreams Environment Sample Project, we developed a pipeline to easily collect aerial and ground views, including ground-truth camera poses and a range of additional data modalities from unmanned aerial vehicle. Our pipeline also provides flexible controls on environmental factors such as light, weather, and types and conditions of wildfire, supporting the need for various tasks covering forest mapping, wildfire emergency response, and beyond. The resulting pilot dataset, WildfireX-SLAM, contains 5.5k low-altitude RGB-D aerial images from a large-scale forest map with a total size of 16 km2. On top of WildfireX-SLAM, a thorough benchmark is also conducted, which not only reveals the unique challenges of 3DGS-based SLAM in the forest but also highlights potential improvements for future works. The dataset and code will be publicly available. Project page: https://zhicongsun.github.io/wildfirexslam.

Figures (7)

• Figure 1: Overview of WildfireX-SLAM. (a) We have constructed a 16 km² forest scene featuring wildfires, offering a high-fidelity forest environment and multi-dimensional wildfire observations. (b) Our framework enables the configuration of various types of wildfires. (c) Through diverse weather and lighting conditions, our framework provides high-fidelity and realistic scenes. (d) We deliver multi-modal data, especially RGB-D from UAVs, to facilitate tasks such as SLAM and navigation.
• Figure 2: Diversity of the WildfireX-SLAM dataset.
• Figure 3: Examples of flight path for data collection in WildfireX-SLAM dataset.
• Figure 4: Mapping results of 3DGS-based SLAM on Challenge Scenario 1 of the WildfireX-SLAM dataset.
• Figure 5: Mapping results of 3DGS-based SLAM on Challenge Scenario 2 of the WildfireX-SLAM dataset.