辽宁工程技术大学学报（自然科学版）2025，Vol.44Issue(3)：257-264,8.DOI:10.11956/j.issn.1008-0562.20240239
改进YOLOv8s的煤矿井下矿工行为检测方法
Improving YOLOv8s for behavior detection of underground miners in coal mine
摘要
Abstract
To address the problem of poor performance in detecting miner behavior in underground coal mine through visual equipment,model optimization and experimental methods are adopted.A benchmark model is constructed based on YOLOv8s.An efficient multi-scale attention mechanism is introduced to improve the backbone network,enhancing the ability to extract and represent multi-pose and multi-scale features of miners.The loss function is optimized to improve the accuracy and stability of low-quality image detection in complex underground scenes;a lightweight module is designed to replace the feature processing module of the original neck deep network,ensuring efficient detection.The results show that the average accuracy of the improved model for underground miner behavior detection is increased by 1.2%,and the amount of model parameters is reduced by 17%.The research conclusions provide reference for the target detection optimization of specific tasks in similar scenarios.关键词
煤矿复杂环境/行为检测/YOLOv8s/注意力机制/轻量化/低质量图像Key words
complex environment of coal mines/behavior detection/YOLOv8s/attention mechanism/lightweight/low quality image分类
矿业与冶金引用本文复制引用
陈伟,穆华星,管彦允,刘珏廷,徐婷婷,王泽华..改进YOLOv8s的煤矿井下矿工行为检测方法[J].辽宁工程技术大学学报（自然科学版）,2025,44(3):257-264,8.基金项目
国家自然科学基金面上项目(52274160 （）
51874300) （）
国家自然科学基金委员会-山西省人民政府煤基低碳联合基金项目(U1510115) （U1510115）
