资源描述:
学校代号 10532 学 号 S1609W0525 分 类 号 TP391 密 级 公开 工程硕士学位论文 基于机器视觉的黑钨矿石智能分选系统研究与 设计 学位申请人姓名 肖继伟 培 养 单 位 电气与信息工程学院 导师姓名及职称 朱青 副教授 童嵘 高级工程师 学 科 专 业 电子与通信工程 研 究 方 向 机器视觉与图像处理 论文提交日期 2019 年 04 月 02 日 万方数据 学校代号10532 学 号S1609W0525 密 级公开 湖南大学工程硕士学位论文 基于机器视觉的黑钨矿石智能分选系统研 究与设计 学位申请人姓名 肖继伟 导师姓名及职称 朱青 副教授 童嵘 高级工程师 培 养 单 位 电气与信息工程学院 专 业 名 称 电子与通信工程 论文提交日期 机器视觉与图像处理 论文答辩日期 2019 年 05 月 12 日 答辩委员会主席 黎福海 教授 万方数据 Research and Design of Black Tungsten Ore Intelligent Sorting System Based on Machine Vision by XIAO Jiwei B.E.Hunan University of Science and Technology2016 A thesis ted in partial satisfaction of the Requirements for the degree of Master of Engineering in Machine Vision and Image Processing in the Graduate School of Hunan University Supervisor Professor ZHU Qing May,2019 万方数据 万方数据 基于机器视觉的黑钨矿石智能分选系统研究与设计 II 摘要 我国钨矿资源丰富,在资源存储量上占有巨大的优势,其中黑钨矿开采价值 较高。目前国内黑钨矿的主要分选方法采用传统的人工分选法,但是,人工分选 法存在效率低、精度低等缺点,且人为主观性强,无法形成客观统一的标准,易造 成资源浪费。近年来,机器视觉技术飞速发展,相较于人工分选方法,具有生产 速率快,效率高且功能稳定等优点,在钨矿工业应用中存在强大优势。利用机器 视觉检测方法代替人工分选不但能消除人为主观因素的存在,而且能形成客观统 一的分选标准,最大限度的提升钨矿的生产效率和矿石分选精度。 本文以黑钨矿为研究对象,分析矿石分选需求,研究并设计一种基于机器视 觉的黑钨矿石智能分选系统,代替传统的人工分选方法,实现黑钨原矿中脉石与 废石的自动化分离。论文主要研究内容如下 首先, 从课题的背景与意义出发, 分析黑钨矿石智能分选系统研究的必要性, 同时引出机器视觉的概念,并介绍机器视觉的系统构成、工业应用和发展现状以 及国内外黑钨矿石分选的研究现状,最终设计本文的整体构架。 其次,设计黑钨矿石智能分选系统的总体结构,其中主要为机械结构、视觉 模块、电气控制模块。根据实际工业需求,分别从矿石传输、视觉成像、矿石分 选三方面设计系统机械结构,接着设计机器视觉模块,包括相机选型、成像方式 等,确保获取高质量的矿石图像,最后为确保系统各功能模块协调工作,设计电 气控制模块,其中包括电气组成框架及其控制流程。 再者,分析黑钨矿石图像的颜色特征,研究黑钨矿石检测算法,主要包括矿 石分割及矿石识别判断两部分,本文中引用遗传神经网络算法实现黑钨矿石图像 的分割,并对分割图像采用连通域分析方法定位脉石位置,通过获取特征区域面 积、长宽比等实现脉石与废石的分类。 最后,设计黑钨矿石智能分选系统软件平台,在 Microsoft Visual Studio 2015 平台使用 C Sharp 语言实现软件开发,最后,针对论文以上的设计,对研究 内容进行总结, 并根据设计中的不足之处作下一步工作展望, 以完善系统的开发。 本文设计的黑钨矿石智能分选系统的各项技术指标矿石分选产量达 40T, 相比人工选矿法增加 33,废石中脉石比率 PT重量比、脉石中废石比率 PN重 量比以及脉石剔除比率 PGT脉石剔除数量比率均符合工业应用的技术指标。 关键词机器视觉;黑钨矿石;遗传神经网络;连通域 万方数据 工程硕士学位论文 III Abstract Chinas tungsten resources are abundant, and it has a huge advantage in resource storage. Among them, the value of black tungsten mining is high. At present, the main sorting of domestic black tungsten ore is traditional manual sorting . However, the manual sorting has the disadvantages of low efficiency and low precision, and people are subjective and cannot objective and unified standards, which is easy to cause waste of resources. . In recent years, machine vision technology has developed rapidly. Compared with manual sorting s, it has the advantages of fast production rate, high efficiency and stable function. It has strong advantages in the application of tungsten ore industry. The use of machine vision detection instead of manual sorting can not only eliminate the existence of human subjective factors, but also an objective and uni sorting standard to maximize the production efficiency of tungsten ore and the accuracy of ore sorting. This paper takes the black tungsten ore as the research object, analyzes the ore sorting demand, researches and designs a black tungsten ore intelligent sorting system based on machine vision, instead of the traditional manual sorting , realizes the gangue and waste rock in the black tungsten ore. Automated separation. The main research contents of the thesis are as follows Firstly, from the background and significance of the subject, the necessity of research on the intelligent sorting system of the black tungsten ore is analyzed. At the same time, the concept of machine vision is introduced, and the system composition, industrial application and development status of machine vision, and black tungsten ore at home and abroad are introduced. The current research status of the selection, and ultimately the overall framework of this paper. Secondly, the overall structure of the black tungsten ore intelligent sorting system is designed, mainly mechanical structure, machine vision module and electrical control module. According to the actual industrial demand, the mechanical structure of the system is designed from the aspects of ore transmission, visual imaging and ore sorting, and then the machine vision module is designed, including camera selection and imaging s, to ensure high quality ore images, and finally to ensure the system. Each functional module coordinates its work and designs an electrical control module that includes an electrical component framework and its control flow. Furthermore, the color characteristics of the black tungsten ore image are analyzed, 万方数据 基于机器视觉的黑钨矿石智能分选系统研究与设计 IV and the black tungsten ore detection algorithm is studied, which mainly includes ore segmentation and ore identification and judgment. In this paper, the genetic neural network algorithm is used to realize the segmentation of the black tungsten ore image, and the segmentation image is adopted. The connected domain analysis locates the location of the gangue and classifies the gangue and waste rock by acquiring features area area, aspect ratio, etc.. Finally, design the black tungsten ore intelligent sorting system software plat, use C Sharp language to realize software development in Microsoft Visual Studio 2015 plat, and finally, summarize the research content for the above paper design, and according to the shortcomings in the design The next step is to improve the development of the system. The technical indicators of the black tungsten ore intelligent sorting system designed in this paper Ore sorting output reaches 40T, which is 33 higher than the manual beneficiation , the ratio of gangue to stone in the waste rock is PT weight ratio, and the ratio of waste rock in the gangue The PN weight ratio and the gangue rejection ratio PGT the number of gangue rejections are in line with the technical specifications of industrial applications. Key words Machine vision; Black tungsten ore; Benetic neural network; Bonnected domain 万方数据 工程硕士学位论文 V 目录 学位论文原创性声明 .................................................................................................. I 学位论文版权使用授权书 .......................................................................................... I 摘要 ............................................................................................................................ II Abstract ................................................................................................................... III 目录 ............................................................................................................................ V 插图索引 ................................................................................................................. VII 插表索引 .................................................................................................................. IX 第 1 章 绪论 ...............................................................................................................1 1.1 研究背景及意义...........................................................................................1 1.2 机器视觉介绍 ..............................................................................................2 机器视觉系统组成 ................................................................................3 机器视觉在工业中的发展现状及应用 .................................................6 1.3 黑钨矿石智能分选系统国内外研究现状 ....................................................7 1.4 课题来源 ......................................................................................................9 1.5 系统需求分析 ............................................................................................ 10 1.6 研究内容与论文构成 ................................................................................. 10 第 2 章 黑钨矿石智能分选系统总体设计 ............................................................... 12 2.1 机械结构设计 ............................................................................................ 12 物料传输装置 ...................................................................................... 12 视觉成像装置 ...................................................................................... 14 矿石分离装置 ...................................................................................... 15 2.2 机器视觉成像方案设计 ............................................................................. 17 工业相机选择 ...................................................................................... 17 相机镜头选择 ...................................................................................... 18 光源及成像方案设计 .......................................................................... 19 2.3 电气控制系统设计 ..................................................................................... 20 电气控制系统组成 .............................................................................. 21 电气控制过程 ...................................................................................... 21 2.4 本章小结 .................................................................................................... 23 第 3 章 黑钨矿石智能分选系统核心算法研究 ....................................................... 24 3.1 黑钨矿石特征分析 ..................................................................................... 24 3.2 黑钨矿石图像分割算法研究 ..................................................................... 26 万方数据 基于机器视觉的黑钨矿石智能分选系统研究与设计 VI 遗传算法 .............................................................................................. 26 神经网络 .............................................................................................. 28 遗传算法优化神经网络 ...................................................................... 33 基于遗传神经网络的黑钨矿石图像分割 ........................................... 36 3.3 黑钨矿石检测算法研究 ............................................................................. 39 基于连通域的矿石定位 ...................................................................... 39 特征分析及矿石判断 .......................................................................... 43 3.4 设备技术指标评估 ..................................................................................... 45 3.5 本章小结 .................................................................................................... 47 第 4 章 黑钨矿石智能分选系统软件设计 ............................................................... 48 4.1 软件系统总体设计 ..................................................................................... 48 4.2 系统软件开发 ............................................................................................ 50 系统软件检测流程 .............................................................................. 50 软件界面 .............................................................................................. 51 软件系统框架介绍 .............................................................................. 53 MYSQL 数据库管理 ............................................................................ 54 4.3 本章小结 .................................................................................................... 56 总结与展望 ............................................................................................................... 57 参考文献 ................................................................................................................... 59 致谢 ........................................................................................................................... 61 附录 A 攻读学位期间参加的科研项目 ................................................................... 62 万方数据 工程硕士学位论文 VII 插图索引 图 1.1 黑钨矿石图 ....................................................................................................1 图 1.2 检测系统组成图 ............................................................................................3 图 1.3 常见视觉系统光学成像方案设计 .................................................................5 图 1.4 机器视觉在各行业的应用 ............................................................................7 图 1.5 TOMRA 分选设备 .........................................................................................8 图 1.6 HPY SORTING 智能选矿机 ..........................................................................9 图 1.7 黑钨矿石分选设备现场图 ............................................................................9 图 2.1 黑钨矿石智能分选系统结构示意图 ........................................................... 12 图 2.2 传送皮带机的功能结构图 .......................................................................... 14 图 2.3 视觉成像装置机械结构图 .......................................................................... 15 图 2.4 高压气枪提出图 .......................................................................................... 15 图 2.5 矿石分离装置实物剔除图 .......................................................................... 16 图 2.6 矿石分离结构示意图 .................................................................................. 17 图 2.7 Dalsa 线阵相机 ............................................................................................ 18 图 2.8 镜头成像示意图 .......................................................................................... 19 图 2.9 成像方案设计图 .......................................................................................... 20 图 2.10 电气控制图 .................................................................................................. 21 图 3.1 黑钨与石英共存 .......................................................................................... 25 图 3.2 石英脉型黑钨矿石 ...................................................................................... 25 图 3.3 粉碎后待检测的原矿 .................................................................................. 25 图 3.4 遗传算法基本流程图 .................................................................................. 28 图 3.5 神经元结构模型图 ...................................................................................... 29 图 3.6 三层 BP 神经网络结构图 ............................................................................ 29 图 3.7 BP 网络的操作步骤图 ................................................................................ 33 图 3.8 遗传算法实数编码序列 .............................................................................. 34 图 3.9 网络优化图 .................................................................................................. 36 图 3.10 算法分割图................................................................................................... 37 图 3.11 图像原图与图像直方图 .............................................................................. 37 图 3.12 矿石图像分割效果图 .................................................................................. 39 图 3.13 像素邻接关系图 .......................................................................................... 40 图 3.14 基于行程的标记方法流程图 ....................................................................... 41 图 3.15 基于种子填充的标记方法 ............................................
展开阅读全文