【作者】 李伟；

【导师】 李鑫； 余玲；

【作者基本信息】 合肥工业大学 ， 控制工程（专业学位）， 2024， 硕士

【摘要】 带式输送机是煤炭开采和运输过程中至关重要的设备。如何对带式输送机进行全面、高效地巡检是保障其安全运行的关键。传统的带式输送机巡检主要采取人工巡检的方式,效率低、查漏率高。因此,利用机器人对煤矿进行智能巡检和监测,提高了煤矿自动化、智能化水平,减轻了巡检人员工作负担,提高了井下安检效率,成为煤矿安全的主要研究方向之一。本文研发一套煤矿井下巷道巡检机器人系统,并提出一种煤流异物检测算法。论文的主要研究工作如下:(1)针对现场煤矿巷道的特殊环境,分析智能巡检的必要性及其具体需求。针对带式输送机所处环境光照条件差、煤流中异物形态多样、轮廓难以检测以及输送带高速移动导致拍摄的图片清晰度低、质量复杂等问题,本文提出了一种基于YOLOv8的带式输送机异物检测方法。先通过预处理来增强目标信息,再添加一种有效的注意力机制,来增强异物的位置信息,同时采用双金字塔结构来应对煤流异物尺度不一的问题,并优化了损失函数提升检测算法的泛化性。本文对从实际矿山收集的数据集进行了模型训练和测试。实验结果表明,本文提出的方法能够有效提高异物检测的准确性,同时缩短训练时间。(2)设计带式输送机智能巡检系统,包括硬件系统和软件系统。硬件系统从控制器、外围硬件选型两个方面进行设计,软件系统从下位机嵌入式软件和上位机后台管理系统两个部分设计。通过实际验证及功能测试,巡检机器人能通过既定的巡检方案获取现场数据并上传,在上位机后台管理系统能实时正确接收并显示获取的数据,并对数据进行分析、处理和统计。更多还原

【Abstract】 Belt conveyor is the most important equipment in coal mining and transportation.How to carry out comprehensive and efficient inspection of belt conveyor is the key to ensure its safe operation.The traditional belt conveyor inspection mainly adopts manual inspection,which has low efficiency and high leak detection rate.Therefore,the use of robots for intelligent inspection and monitoring of coal mines has become one of the main research directions of coal mine safety,which has improved the automation and intelligence level of coal mines,reduced the work burden of inspection personnel,and improved the efficiency of underground security check.In this paper,a set of inspection robot system for underground roadway in coal mine is developed,and an algorithm for detecting foreign objects in coal flow is proposed.The main research work of this paper is as follows:(1)According to the special environment of coal mine roadway on site,analyze the necessity of intelligent inspection and its specific needs.Aiming at such problems as poor lighting conditions in the environment of belt conveyor,diverse forms of foreign objects in coal flow,difficult to detect contours,and low clarity and complex quality of pictures taken due to high-speed movement of conveyor belt,this paper proposes a YOLOv8 based foreign objects detection method for belt conveyor.First,the target information is enhanced by preprocessing,and then an effective attention mechanism is added to enhance the location information of foreign bodies.At the same time,the double pyramid structure is adopted to deal with the problem of different scales of foreign bodies in coal flow,and the loss function is modified.We trained and tested the model on data sets collected from actual mines.The experimental results show that the proposed method can effectively improve the accuracy of foreign body detection and shorten the training time.(2)Design intelligent inspection system for belt conveyor,including hardware system and software system.The hardware system is designed from the two aspects of controller and peripheral hardware selection,and the software system is designed from the embedded software of the lower computer and the monitoring software of the upper computer.Through practical verification and functional testing,the inspection robot can obtain and upload on-site data through the established inspection scheme,and the monitoring software on the upper computer can receive and display the obtained data correctly in real time,and analyze,process and make statistics on the data.更多还原