Real-Time Aerial Fire Detection on Resource-Constrained Devices Using Knowledge Distillation

TL;DR

This paper introduces a lightweight fire detection model using knowledge distillation from a stronger teacher model, enabling real-time wildfire detection on resource-constrained devices like UAVs and IoT sensors.

Contribution

The work presents a novel, efficient fire detection model based on MobileViT-S, compressed via knowledge distillation, suitable for deployment on edge devices for early wildfire detection.

Findings

Achieves real-time fire detection on resource-limited devices.

Outperforms state-of-the-art models in accuracy by 0.44% and 2.00%.

Maintains high detection accuracy with a significantly smaller model size.

Abstract

Wildfire catastrophes cause significant environmental degradation, human losses, and financial damage. To mitigate these severe impacts, early fire detection and warning systems are crucial. Current systems rely primarily on fixed CCTV cameras with a limited field of view, restricting their effectiveness in large outdoor environments. The fusion of intelligent fire detection with remote sensing improves coverage and mobility, enabling monitoring in remote and challenging areas. Existing approaches predominantly utilize convolutional neural networks and vision transformer models. While these architectures provide high accuracy in fire detection, their computational complexity limits real-time performance on edge devices such as UAVs. In our work, we present a lightweight fire detection model based on MobileViT-S, compressed through the distillation of knowledge from a stronger teacher model. The ablation study highlights the impact of a teacher model and the chosen distillation technique on the model's performance improvement. We generate activation map visualizations using Grad-CAM to confirm the model's ability to focus on relevant fire regions. The high accuracy and efficiency of the proposed model make it well-suited for deployment on satellites, UAVs, and IoT devices for effective fire detection. Experiments on common fire benchmarks demonstrate that our model suppresses the state-of-the-art model by 0.44%, 2.00% while maintaining a compact model size. Our model delivers the highest processing speed among existing works, achieving real-time performance on resource-constrained devices.