Abstract: Aiming to address issues such as confusion between smoke flame targets and background within complex backgrounds, which often result in low accuracy of smoke flame detection, an enhanced model based on YOLOv8s for detecting smoke flames within complex backgrounds is proposed. Firstly, the feature channels are highly similar to each other, and in order to effectively utilize the redundancy across different channels and improve the  ability  of model  to differentiate between smoke and flame targets and backgrounds, the C2fFR (C2f with partial rep conv) lightweight feature extraction module is introduced. Secondly, the MCFM (multi-scale context fusion module) is designed to capture and utilize contextual information for enhancing feature representation. Lastly, the Inner-SIoU loss function is employed to address bounding box mismatches and the  regression ability of the model is improved for high IoU samples. Experimental results demonstrate that compared to the baseline YOLOv8s model, the enhanced YOLOv8s smoke flame detection model achieves improvements of 4.6 percentage points in mAP@50 and 2.3 percentage points in mAP@50:95. Moreover, it reduces the number of model parameters by 18.9% and computation by 8.1%.  while maintaining an FPS (frame per second) of 93. Additionally, it exhibits superior detection performance when compared to other mainstream detection algorithms.

Key words: YOLOv8s, C2fFR, multi-scale context fusion module, Inner-SIoU

摘要： 针对复杂背景下烟雾火焰目标与背景混淆，导致烟雾火焰检测精度低等问题，提出一种面向复杂背景下烟雾火焰检测的YOLOv8s改进模型。特征通道之间具有高度相似性，为了有效利用跨不同通道间的冗余，提高模型对烟雾火焰目标和背景的区分，设计了C2fFR（C2f with partial rep conv）轻量级特征提取模块。设计了MCFM（multi-scale context fusion module）多尺度上下文融合模块，来捕捉并利用上下文信息，增强特征的表示。使用Inner-SIoU损失函数，解决边界框不匹配的问题，提高模型对高IoU样本的回归能力。实验结果表明，改进后的YOLOv8s烟雾火焰检测模型相比于基线模型YOLOv8s，mAP@50提升了4.6个百分点，mAP@50：95提升了2.3个百分点，模型参数量降低了18.9%，计算量降低了8.1%，FPS为93帧/s，与其他主流检测算法相比也具有较好的检测性能。

关键词: YOLOv8s, C2fFR, 多尺度上下文融合, Inner-SIoU