摘要
针对目前煤矿在运输过程中皮带保护系统响应滞后、人工依赖度高的现状,容易发生落料口堵料、皮带偏料溢料等问题。提出了一种基于多模态感知的AI皮带运输煤矿安全监控系统设计方案,通过部署工业摄像头与线结构光设备,构建物联网感知层实时采集皮带运行视频流与物料轮廓数据;在智能分析层融合可见光与热成像双光谱信息,结合光流法与YOLOv5目标检测算法,实现对皮带跑偏、堵料、溢料及托辊偏移等常见故障的实时监测和异常检测;异常数据处理层通过PLC联动控制实现预警显示或自动停机保护,形成“感知-分析-响应”闭环管控机制。相比于现有的研究方案,该方案不仅降低了人工成本,而且为复杂工业场景下的输送带智能监测提供了可复用的技术框架,提高了煤矿生产的安全性和效率。
To address the current issues of delayed response in conveyor belt protection systems and high reliance on manual monitoring during coal mine transportation,which frequently lead to material blockage at discharge points,belt misalignment,and overflow,this paper proposes a multi-modal sensing-based AI-driven coal mine safety monitoring system for belt transportation.The system deploys industrial cameras and line-structured light devices to construct an IoT perception layer for real-time acquisition of belt operation video streams and material contour data.At the intelligent analysis layer,visible light and thermal imaging dual-spectral information are fused,integrating optical flow analysis and the YOLOv5 object detection algorithm to achieve real-time monitoring and anomaly detection of common faults such as belt deviation,material blockage,overflow,and roller displacement.The anomaly processing layer implements warning displays and automatic shutdown protection through PLC linkage control,forming a closed-loop"perception-analysis-response"management mechanism.Compared with existing solutions,the proposed system not only reduces labor costs but also provides a reusable technical framework for intelligent conveyor belt monitoring in complex industrial scenarios,thereby enhancing both the safety and efficiency of coal production.
作者
康文坚
龚国义
任伟业
Kang Wenjian;Gong Guoyi;Ren Weiye(Fujian Huadian Storage and Transportation Co.,Ltd.,Fuzhou Fujian 350512,China)
出处
《机械管理开发》
2025年第11期266-270,共5页
Mechanical Management and Development
基金
福建省自然科学基金面上项目(2022J01551)。
关键词
皮带运输
人工智能
物联网
异常检测
belt conveyor transportation
artificial intelligence
Internet of Things
anomaly detection
