煤矿安全2025，Vol.56Issue(5)：236-244,9.DOI:10.13347/j.cnki.mkaq.20241343
基于多通道图像处理的斜巷绞车运输闭锁系统
Ramp winch locking system based on multi-channel image processing
摘要
Abstract
Ramp winch locking system is an important means of avoiding the simultaneous operation of winches and people walking in the ramp.The traditional locking system has the problems of low accuracy,difficult installation and maintenance,and high hard-ware cost,and for these problems,a ramp winch locking system based on multi-channel image processing is proposed,and an in-trinsically safe locking dection system is developed.The algorithms are categorized into object detection,tracking and occlusion de-tection.In the object detection,Slim-C3 was designed using GSConv to lighten the backbone and neck networks and to improve the accuracy of target detection by the decoupling head based on GSConv;in the target tracking stage,the tracking algorithm of second-level matching is studied to improve the speed without decreasing the accuracy.The experimental results show that the proposed ob-ject detection optimized the AP0.5:0.95 for people and winch by 0.57%and 1.65%over YOLOv8s,respectively;the proposed tracking improved by 0.55%over YOLOv8s+DeepSort in IDF1;the precision and recall rates of the count were 96.43%and 93.10%,respect-ively.关键词
煤矿智能化/煤矿运输系统/绞车运输/闭锁系统/目标检测/目标跟踪/计算机视觉Key words
coal mine intelligentization/coal mine transportation system/winch transportation/locking system/object detection/object tracking/computer vision分类
矿业与冶金引用本文复制引用
游磊,李一文,张海峰,杨筱彧,罗明华..基于多通道图像处理的斜巷绞车运输闭锁系统[J].煤矿安全,2025,56(5):236-244,9.基金项目
天地科技股份有限公司科技创新创业资金专项资助项目(2024-TD-MS007) （2024-TD-MS007）
