Advancing Real-Time Aerial Wildfire Detection Through Plume Recognition and Knowledge Distillation
Open Access
Abstract
Uncrewed aerial systems (UAS)-based remote sensing and artificial intelligence (AI) analysis enable real-time wildfire or bushfire detection, facilitating early response to minimize damage and protect lives and property. However, their effectiveness is limited by three issues: distinguishing smoke from fog, the high cost of manual annotation, and the computational demands of large models. This study addresses the three key challenges by introducing plume as a new indicator to better distinguish smoke from similar visual elements, and by employing a hybrid annotation method using knowledge distillation (KD) to reduce expert labour and accelerate labelling. Additionally, it leverages lightweight YOLO Nano models trained with pseudo-labels generated from a fine-tuned teacher network to lower computational demands while maintaining high detection accuracy for real-time wildfire monitoring. Controlled pile burns in Canungra, QLD, Australia, were conducted to collect UAS-captured images over deciduous vegetation, which were subsequently augmented with the Flame2 dataset, which contains wildfire images of coniferous vegetation. A Grounding DINO model, fine-tuned using few-shot learning, served as the teacher network to generate pseudo-labels for a significant portion of the Flame2 dataset. These pseudo-labels were then used to train student networks consisting of YOLO Nano architectures, specifically versions 5, 8, and 11 (YOLOv5n, YOLOv8n, YOLOv11n). The experimental results show that YOLOv8n and YOLOv5n achieved an mAP@0.5 of 0.721. Plume detection outperforms smoke indicators (F1: 76.1–85.7% vs. 70%) in fog and wildfire scenarios. These findings underscore the value of incorporating plume as a distinct class and utilizing KD, both of which enhance detection accuracy and scalability, ultimately supporting more reliable and timelier wildfire monitoring and response.Keywords
Funding Information
- Australian Research Council (DP 220103233)
This publication has 38 references indexed in Scilit:
- A survey on knowledge distillation: Recent advancementsMachine Learning with Applications, 2024
- Transformer models for Land Cover Classification with Satellite Image Time SeriesPFG – Journal of Photogrammetry, Remote Sensing and Geoinformation Science, 2024
- Local Evaluation of Large-scale Remote Sensing Machine Learning-generated Building and Road Dataset: The Case of RwandaPFG – Journal of Photogrammetry, Remote Sensing and Geoinformation Science, 2024
- Automatic labelling framework for optical remote sensing object detection samples in a wide area using deep learningExpert Systems with Applications, 2024
- DUDES: Deep Uncertainty Distillation using Ensembles for Semantic SegmentationPFG – Journal of Photogrammetry, Remote Sensing and Geoinformation Science, 2024
- Semantic Real-Time Mapping with UAVsPFG – Journal of Photogrammetry, Remote Sensing and Geoinformation Science, 2023
- The influence of wind on smoke propagation to the lower layer in naturally ventilated tunnelsTunnelling and Underground Space Technology, 2022
- A review of UAV platforms, sensors, and applications for monitoring of sugarcane cropsRemote Sensing Applications: Society and Environment, 2022
- A real-time deep learning forest fire monitoring algorithm based on an improved Pruned + KD modelJournal of Real-Time Image Processing, 2021
- Video Fire Smoke Detection Using Motion and Color FeaturesFire Technology, 2009