TF-1770: A Visible–Infrared Forest Fire Image Dataset for Multimodal Perception and Image Fusion Research

TF-1770 is a multimodal visible–infrared image dataset designed for research on forest fire monitoring, multimodal perception, and visible–infrared image fusion in complex wildfire environments. The dataset contains 1,770 spatially aligned pairs of visible and thermal infrared images, collected from both unmanned aerial vehicle (UAV) platforms and ground-based firefighting perspectives. Forest fire monitoring often involves challenging environmental conditions, including dense smoke, variable illumination, complex vegetation backgrounds, and long observation distances. In such scenarios, visible and infrared imaging provide complementary information. Visible images capture detailed structural and texture information of the scene, while infrared images are highly sensitive to thermal targets such as flames and high-temperature regions. The TF-1770 dataset was constructed to facilitate research on effectively integrating these complementary modalities for improved wildfire perception and monitoring. To better reflect real-world wildfire monitoring conditions, the dataset was collected using multiple sensing platforms representing different observation perspectives. The UAV subset contains aerial observations acquired from unmanned aerial platforms at different flight altitudes, which provide large-scale views of wildfire spread patterns and thermal distribution. The firefighter subset contains ground-level observations captured from sensing devices simulating the operational perspectives of firefighters during field operations. All image pairs in TF-1770 are spatially registered, enabling direct use for multimodal perception tasks such as image fusion and cross-modal feature learning. Dataset Composition The TF-1770 dataset contains a total of 1,770 aligned visible–infrared image pairs, organized according to acquisition perspectives: UAV viewpoint: 944 image pairs Captured from aerial platforms at different observation heights Provide broader scene coverage and large-scale fire spread information Firefighter viewpoint: 826 image pairs Captured from ground-level sensing devices simulating firefighter operational perspectives Contain closer observations of flames, smoke, and complex forest structures Each visible image has a corresponding infrared image with the same filename, ensuring straightforward pairing for multimodal tasks. Data Acquisition Data were collected in representative forest environments using a multi-platform sensing strategy. Aerial observations were acquired using UAV platforms to capture wildfire scenes at different spatial scales and observation distances. Ground-level observations were obtained using dual-spectrum sensing devices positioned at viewpoints similar to those encountered by firefighters during field operations. This hierarchical acquisition strategy enables the dataset to capture variations in target scale, observation distance, and scene complexity, which are common in real wildfire monitoring scenarios. Dataset Applications The TF-1770 dataset supports research in several areas related to wildfire monitoring and multimodal perception, including: Visible–infrared image fusion Multimodal wildfire detection Cross-modal feature learning Edge intelligent perception systems Ecological monitoring and disaster response technologies Dataset Structure The dataset is organized according to imaging modality and acquisition perspective: TF-1770├── visible│ ├── uav│ └── firefighter│├── infrared│ ├── uav│ └── firefighter Each visible image has a corresponding infrared image with the same filename. Citation If you use the TF-1770 dataset in your research, please cite the following work: Qian, Y., Zhang, G., et al.FF-Fusion: A Teacher–Student Distillation Framework for Lightweight Visible–Infrared Image Fusion in Forest Fire Scenarios. License The dataset is released for academic research purposes only.Users are encouraged to cite the dataset and the associated publication when using the data in research. Contact For questions regarding the dataset, please contact: Yu Qian