Fire-YOLO: A Small Target Object Detection Method for Fire Inspection

TLDR

The study proposes an improved Fire‑YOLO deep‑learning algorithm to detect small fire‑like and smoke‑like targets in forest fire images under varying natural lighting, and to extend this approach to other complex small‑target detection scenarios. Fire‑YOLO expands its feature‑extraction network to three dimensions, enhances feature propagation for small‑target identification, reduces model parameters, and employs a feature‑pyramid strategy to select the top‑performing prediction box. Fire‑YOLO outperforms state‑of‑the‑art detectors in locating small fire and smoke targets, effectively inspects such objects, and achieves real‑time performance with an average inference time of 0.04 s per 416 × 416 frame.

Abstract

For the detection of small targets, fire-like and smoke-like targets in forest fire images, as well as fire detection under different natural lights, an improved Fire-YOLO deep learning algorithm is proposed. The Fire-YOLO detection model expands the feature extraction network from three dimensions, which enhances feature propagation of fire small targets identification, improves network performance, and reduces model parameters. Furthermore, through the promotion of the feature pyramid, the top-performing prediction box is obtained. Fire-YOLO attains excellent results compared to state-of-the-art object detection networks, notably in the detection of small targets of fire and smoke. Overall, the Fire-YOLO detection model can effectively deal with the inspection of small fire targets, as well as fire-like and smoke-like objects. When the input image size is 416 × 416 resolution, the average detection time is 0.04 s per frame, which can provide real-time forest fire detection. Moreover, the algorithm proposed in this paper can also be applied to small target detection under other complicated situations.