FireRescue: A UAV-Based Dataset and Enhanced YOLO Model for Object Detection in Fire Rescue Scenes

TL;DR

FireRescue tackles the lack of domain-specific benchmarks for aerial fire-rescue perception and targets critical urban-rescue elements such as fire trucks, firefighters, flames, and smoke. The authors introduce FireRescue, a large UAV-based dataset with 15980 images and 32000 annotations across 8 categories, and propose FRS-YOLO, an enhanced one-stage detector featuring a Multi-Dimensional Collaborative Enhancement Attention (MCEA) module and the Dysample dynamic upsampling method. Across nano- and small-scale variants, FRS-YOLO delivers consistent improvements in $mAP_{50}$, $mAP_{50-95}$, Precision, and Recall over YOLOv12 baselines, while maintaining lightweight inference. Visualization analyses using detection results and Grad-CAM heatmaps confirm more focused attention on actionable targets and reduced category confusion in cluttered fire-rescue scenes. Overall, the work provides a practical, open benchmark and a robust, efficient detector to support real-time fire-rescue command and decision-making.

Abstract

Object detection in fire rescue scenarios is importance for command and decision-making in firefighting operations. However, existing research still suffers from two main limitations. First, current work predominantly focuses on environments such as mountainous or forest areas, while paying insufficient attention to urban rescue scenes, which are more frequent and structurally complex. Second, existing detection systems include a limited number of classes, such as flames and smoke, and lack a comprehensive system covering key targets crucial for command decisions, such as fire trucks and firefighters. To address the above issues, this paper first constructs a new dataset named "FireRescue" for rescue command, which covers multiple rescue scenarios, including urban, mountainous, forest, and water areas, and contains eight key categories such as fire trucks and firefighters, with a total of 15,980 images and 32,000 bounding boxes. Secondly, to tackle the problems of inter-class confusion and missed detection of small targets caused by chaotic scenes, diverse targets, and long-distance shooting, this paper proposes an improved model named FRS-YOLO. On the one hand, the model introduces a plug-and-play multidi-mensional collaborative enhancement attention module, which enhances the discriminative representation of easily confused categories (e.g., fire trucks vs. ordinary trucks) through cross-dimensional feature interaction. On the other hand, it integrates a dynamic feature sampler to strengthen high-response foreground features, thereby mitigating the effects of smoke occlusion and background interference. Experimental results demonstrate that object detection in fire rescue scenarios is highly challenging, and the proposed method effectively improves the detection performance of YOLO series models in this context.

Figures (15)

• Figure 1: Annotation Statistics of the FireRescue Dataset
• Figure 2: Structure of FRS-YOLO
• Figure 3: Structure of MCEA
• Figure 4: Structure of Triple Squeeze Transformation
• Figure 5: Sampling based dynamic upsampling and module designs in DySample