基于YOLO-MSD 的矿井烟雾检测模型

金属矿山 ›› 2026, Vol. 55 ›› Issue (2): 218-228.

基于YOLO-MSD 的矿井烟雾检测模型

贾时东　王静宇　任国印　任鹏举

内蒙古科技大学数智产业学院,内蒙古包头市 014010

出版日期:2026-02-15 发布日期:2026-03-04
通讯作者: 王静宇(1976—),男,教授,博士,硕士研究生导师。
作者简介:贾时东(1999—),男,硕士研究生。
基金资助:
国家自然科学基金项目(编号:62466045);内蒙古自治区重点研发和成果转化计划项目(编号:2022YFSH0044);内蒙古自治区自然科
学基金项目(编号:2025MS06018)。

Mine Smoke Detection Model Based on YOLO-MSD

JIA Shidong　WANG Jingyu　REN Guoyin　REN Pengju

School of Cyber Science and Technology,Inner Mongolia University of Science & Technology,Baotou 014010,China

Online:2026-02-15 Published:2026-03-04

摘要/Abstract

摘要： 针对矿井烟雾检测存在粉尘、金属反光、光照不均、目标尺度不一致等复杂情况,导致检测速度慢、精度
不高的问题,提出一种改进YOLOv8n 的矿井烟雾检测模型YOLO-MSD。首先,在主干网络中引入RepVGGBlock 以提
升复杂干扰下的检测表现。其次,设计了共享卷积金字塔(FPSC)替代SPPF,有效提升多尺度特征感知能力并降低计
算量。最后,采用BiFPN 的加权融合思想改进多分枝辅助特征金字塔网络(MAFPN)作为网络结构,通过设计浅层辅
助加权融合(SAWF)和深层辅助加权融合(AAWF)模块,在提升小目标烟雾检测精度的同时有效减少参数量。试验
结果表明,YOLO-MSD 在保持255 帧/ s 推理速度的同时,达到91. 7%的AP50 与74. 5%的AP50-95,较YOLOv8n 分别
提升4. 0 和6. 1 个百分点,在复杂的矿井环境下仍能保持稳定的烟雾检测性能。

关键词: 矿井烟雾检测　RepVGGBlock　YOLOv8n　多尺度特征融合　MAFPN

Abstract: Aiming at the complex situation of dust,metal reflection,uneven illumination and inconsistent target scale in
mine smoke detection,which leads to slow detection speed and low accuracy,a mine smoke detection model YOLO-MSD based
on improved YOLOv8 n is proposed. Firstly,RepVGGBlock is introduced into the backbone network to improve the detection
performance under complex interference. Secondly,a shared convolutional pyramid (FPSC) is designed to replace SPPF,which
effectively improves the multi-scale feature perception ability and reduces the amount of calculation. Finally,the weighted fusion
idea of BiFPN is used to improve the multi-branch auxiliary feature pyramid network (MAFPN) as the network structure.
By designing shallow auxiliary weighted fusion (SAWF) and deep auxiliary weighted fusion (AAWF) modules,the accuracy of
small target smoke detection is improved and the number of parameters is effectively reduced. The experimental results show
that YOLO-MSD achieves 91. 7% AP50 and 74. 5% AP50-95 while maintaining the reasoning speed of 255 FPS,which is 4. 0
and 6. 1 percent points higher than that of YOLOv8n,respectively. It can still maintain stable smoke detection performance in
complex mine environment.

Key words: mine smoke detection,RepVGGBlock,YOLOv8n,multi-scale feature fusion,MAFPN

中图分类号:

TD752

贾时东　王静宇　任国印　任鹏举. 基于YOLO-MSD 的矿井烟雾检测模型[J]. 金属矿山, 2026, 55(2): 218-228.

JIA Shidong　WANG Jingyu　REN Guoyin　REN Pengju. Mine Smoke Detection Model Based on YOLO-MSD[J]. Metal Mine, 2026, 55(2): 218-228.

扫码分享