|本期目录/Table of Contents|

[1]司聿宣,范韬.煤矿井下火灾多尺度视觉智能识别方法研究[J].中国安全生产科学技术,2026,22(5):158-166.[doi:10.11731/j.issn.1673-193x.2026.05.019]
　Si Yuxuan,Fan Tao.Study on multi-scale visual intelligence-based recognition methods for underground coal mine fires[J].Journal of Safety Science and Technology,2026,22(5):158-166.[doi:10.11731/j.issn.1673-193x.2026.05.019]
点击复制

煤矿井下火灾多尺度视觉智能识别方法研究

《中国安全生产科学技术》[ISSN:1673-193X/CN:11-5335/TB]

卷:
22
期数:
2026年5期
页码:
158-166
栏目:
火灾与爆炸安全
出版日期:
2026-05-30

文章信息/Info

Title:
Study on multi-scale visual intelligence-based recognition methods for underground coal mine fires
文章编号:
1673-193X(2026)-05-0158-09
作者:
司聿宣; 范韬
（1.烟台职业学院，山东 烟台 264670；
2.中国安全生产科学研究院，北京 100012）
Author(s):
Si Yuxuan;  Fan Tao
（1.Yantai Vocational College,Yantai Shandong 264670,China;
2.China Academy of Safety Science and Technology,Beijing 100012,China）
关键词:
多尺度检测; 矿井火灾; 注意力机制; 目标检测
Keywords:
multi-scale detection;  coal mine fire;  attention mechanism;  target detection
分类号:
X932
DOI:
10.11731/j.issn.1673-193x.2026.05.019
文献标志码:
文献标志码：A
摘要:
为解决复杂环境下矿井传统火灾检测算法对火灾小目标区域易漏检的问题，提出1种基于YOLOv3网络改进的井下火灾检测方法，实现深浅层特征的有效融合，并将原有三尺度预测扩展为四尺度预测，提升小目标火灾区域的检测能力。研究结果表明：与原始YOLOv3算法相比，改进算法的平均检测精度提高6.6百分点，召回率提升5.8百分点；相较于YOLOv5等算法，在复杂矿井环境下表现出更优的检测性能。该方法在保证实时性的同时显著提高检测精度，可有效增强井下火灾识别能力。研究结果对矿井火灾早期预警与安全管理具有重要的理论意义和工程应用价值。
Abstract:
In order to address the issue of traditional fire detection algorithms in mines under complex environments being prone to missing small fire target areas,an improved underground fire detection method based on the YOLOv3 network is proposed.This method effectively integrates deep and shallow features and expands the original three-scale prediction to four-scale prediction,thereby enhancing the detection capability of small fire target areas.The research results show that compared with the original YOLOv3 algorithm,the average detection accuracy of the improved algorithm has increased by 6.6 percentage points,and the recall rate has improved by 5.8 percentage points.Compared with algorithms such as YOLOv5,it demonstrates superior detection performance in complex mine environments.This method significantly improves detection accuracy while ensuring real-time performance,effectively enhancing the ability to identify underground fires and has important theoretical significance and engineering application value for early warning and safety management of coal mine fires.

参考文献/References:

［1］李正周,方朝阳,顾园山,等.基于无线多传感器信息融合的火灾检测系统［J］.数据采集与处理,2014,29(5):694-698. Li Zhengzhou,Fang Chaoyang,Gu Yuanshan,et al.Fire detection system based on wireless multi-sensor information fusion［J］.Journal of Data Acquisition and Processing,2014,29(5):694-698.
［2］Michalski S R.The Jharia mine fire control technical assistance project:an analysis［J］.International Journal of Coal Geology,2004,59(1-2):83-90.
［3］叶铭亮,周慧英,李建军.基于改进Swin Transformer的森林火灾检测算法［J］.中南林业科技大学学报,2022(8):101-110. Ye Mingliang,Zhou Huiying,Li Jianjun.Forest fire detection algorithm based on an improved swin transformer［J］.Journal of Central South University of Forestry & Technology,2022(8):101-110.
［4］刘洲岐,王雷,刘聪,等.一种基于颜色模型的火灾识别系统［J］.山东理工大学学报(自然科学版),2022,36(3):1-6. Liu Zhouqi,Wang Lei,Liu Cong,et al.A fire identification system based on the color model［J］.Journal of Shandong University of Technology (Natural Science Edition),2022,36(3):1-6.
［5］卢鹏,赵亚琴,陈越,等.复杂背景环境下基于SSD_MobileNet深度学习模型的火焰图像识别研究［J］.火灾科学,2020,29(3):142-149. Lu Peng,Zhao Yaqin,Chen Yue,et al.Flame recognition in complex environment based on deep learning model SSD_MobileNet［J］.Fire Science,2020,29(3):142-149.
［6］王伟峰,张宝宝,王志强,等.基于YOLOv5的矿井火灾视频图像智能识别方法［J］.工矿自动化,2021,47(9):53-57. Wang Weifeng,Zhang Baobao,Wang Zhiqiang,at al.Intelligent recognition method of mine fire video image based on YOLOv5［J］.Industry and Mine Automation,201,47(9):53-57.
［7］Zhang Li,Zhu Yuqin,Wu Hao,et al.An optimized multisource bilinear convolutional neural network model for flame image identification of coal mine［J］.IEEE Access,2022,10:47284-47300.
［8］孙婷婷.基于静态特征和动态行为的火灾检测模型［J］.中国安全生产科学技术,2021,17(1):96-101. Sun Tingting.Fire detection model based on static characteristics and dynamic behavior［J］.Journal of Safety Science and Technology,2021,17(1):96-101.
［9］邓力,刘全义,胡林,等.基于BP神经网络的受限空间火灾联合探测方法［J］.中国安全生产科学技术,2020,16(1):158-162. Deng Li,Liu Quanyi,Hu Lin,at al.Joint detection method of fire in confined space based on BP neural network［J］.Journal of Safety Science and Technology,2020,16(1):158-162.
［10］Farhadi A,Redmon J.Yolov3:an incremental improvement［C/OL］//Computer vision and pattern recognition.Berlin/Heidelberg,Germany:Springer,2018,1804:1-6.https://arxiv.org/abs/1804.02767.
［11］Kong Derui,Li Yinfeng,Duan Manzhen.Fire and smoke real-time detection algorithm for coal mines based on improved YOLOv8s［J］.Plos One,2024,19(4):e0300502.
［12］Yang Yingbing,Zhao Duan,Ge Yicheng,et al.Design and research of an intelligent detection method for coal mine fire edges［J］.Applied Sciences,2025,15(19):10589.
［13］Lin T Y,Dollr P,Girshick R,et al.Feature pyramid networks for object detection［C/OL］//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,2017:936-944.https://ieeexplore.ieee.org/document/8099589.DOI:10.1109/CVPR.2017.106.
［14］Redmon J,Farhadi A.YOLO9000:better,faster,stronger［C/OL］//Proceedings of the IEEE conference on computer vision and pattern recognition,2017:6517-6525.https://ieeexplore.ieee.org/document/8100173.DOI:10.1109/CVPR.2017.690.
［15］He Kaiming,Sun Jian,Tang Xiaoou.Single image haze removal using dark channel prior［J］.IEEE Transactions on Pattern Analysis and Machine Intelligence,2010,33(12):2341-2353.
［16］Zhu Linxianga,Lee F,Cai Jiaweia,et al.An improved feature pyramid network for object detection［J］.Neurocomputing,2022,483:127-139.
［17］Ren Shaoqing,He Kaiming,Girshick R,et al.Faster R-CNN:towards real-time object detection with region proposal networks［J］.IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,39(6):1137-1149.
［18］Liu Wei,Anguelov D,Erhan D,et al.SSD:single shot multibox detector［C/OL］//European Conference on Computer Vision.Cham:Springer International Publishing,2016:21-37.https://link.springer.com/chapter/10.1007/978-3-319-46448-0_2#citeas.DOI:10.1007/978-3-319-46448-0_2.

相似文献/References:

[1]张龙,袁树杰,朱成涛.俯采工作面采空区“三带”分布研究[J].中国安全生产科学技术,2013,9(5):67.[doi:10.11731/j.issn.1673-193x.2013.05.013]
　ZHANG Long,YUAN Shu jie,ZHU Cheng tao,et al.Study on distribution of “Three Zones” in goaf of fully mechanized underhand mining face[J].Journal of Safety Science and Technology,2013,9(5):67.[doi:10.11731/j.issn.1673-193x.2013.05.013]
[2]邬云龙,谭波.基于粗糙集Skowron差别矩阵的矿井火灾风险评价指标约简[J].中国安全生产科学技术,2016,12(5):60.[doi:10.11731/j.issn.1673-193x.2016.05.011]
　WU Yunlong,TAN Bo.Reduction of risk assessment indexes for mine fire based on Skowron discernibility matrix of rough set theory[J].Journal of Safety Science and Technology,2016,12(5):60.[doi:10.11731/j.issn.1673-193x.2016.05.011]
[3]贾静,郭立稳,朱令起,等.矿井巷道火灾烟流逆退数值模拟及临界风速研究[J].中国安全生产科学技术,2020,16(4):94.[doi:10.11731/j.issn.1673-193x.2020.04.015]
　JIA Jing,GUO Liwen,ZHU Lingqi,et al.Study on numerical simulation of smoke backflow and critical wind speed in mine roadway fire[J].Journal of Safety Science and Technology,2020,16(5):94.[doi:10.11731/j.issn.1673-193x.2020.04.015]
[4]胡纪年,李雨成,李俊桥,等.基于CNN的矿井外因火灾火源定位方法研究[J].中国安全生产科学技术,2024,20(3):134.[doi:10.11731/j.issn.1673-193x.2024.03.019]
　HU Jinian,LI Yucheng,LI Junqiao,et al.Study on localization method of mine exogenous fire source based on CNN[J].Journal of Safety Science and Technology,2024,20(5):134.[doi:10.11731/j.issn.1673-193x.2024.03.019]
[5]闫帅,王刚,洪林,等.细水雾与纵向通风对倾斜巷道火灾烟气运动规律影响研究*[J].中国安全生产科学技术,2024,20(2):188.[doi:10.11731/j.issn.1673-193x.2024.02.026]
　YAN Shuai,WANG Gang,HONG Lin,et al.Research on influence of water mist and longitudinal ventilation on smoke movement in inclined roadway fire[J].Journal of Safety Science and Technology,2024,20(5):188.[doi:10.11731/j.issn.1673-193x.2024.02.026]
[6]李贺,田丽,曾钢,等.基于FDS的风速对矿井火灾蔓延规律的影响研究*[J].中国安全生产科学技术,2022,18(5):143.[doi:10.11731/j.issn.1673-193x.2022.05.022]
　LI He,TIAN Li,ZENG Gang,et al.Study on influence of wind speed on spread law of mine fire based on FDS[J].Journal of Safety Science and Technology,2022,18(5):143.[doi:10.11731/j.issn.1673-193x.2022.05.022]

备注/Memo

备注/Memo:
收稿日期： 2026-02-04；修回日期：2026-03-23；网络首发日期：2026-05-09
作者简介： 司聿宣，硕士，讲师，主要研究方向为计算机建模与安全管理。
通信作者： 范韬，博士，工程师，主要研究方向为合成材料、锂电池安全。
更新日期/Last Update: 2026-06-03