img

QQ群聊

img

官方微信

高级检索

黄金科学技术 ›› 2022, Vol. 30 ›› Issue (1): 131-140.doi: 10.11872/j.issn.1005-2518.2022.01.135

• 采选技术与矿山管理 • 上一篇    下一篇

基于图像识别的井下机车载矿量计量系统

张钰鹏(),吴富姬,郭毅   

  1. 赣州有色冶金研究所有限公司,有色金属矿冶装备工作设计中心,江西 赣州 341000
  • 收稿日期:2021-09-26 修回日期:2022-01-10 出版日期:2022-02-28 发布日期:2022-04-25
  • 作者简介:张钰鹏(1993-),男,江西进贤人,工程师,从事软件开发和自动化技术研究工作。18870887950@163.com
  • 基金资助:
    江西省重点研发计划项目“基于图像识别和‘大数据’的非接触式运矿计量系统成套技术及装备研发”(20202BBEL53016)

Mine-Loading Measurement System of Underground Locomotive Based on Image Recognition

Yupeng ZHANG(),Fuji WU,Yi GUO   

  1. Nonferrous Metal Mining and Metallurgy Equipment Work Design Center,Ganzhou Nonferrous Metallurgy Research Institute Co. ,Ltd. ,Ganzhou 341000,Jiangxi,China
  • Received:2021-09-26 Revised:2022-01-10 Online:2022-02-28 Published:2022-04-25

摘要:

出矿量是矿山企业制定生产计划的一个重要指标,而目前绝大多数地下矿山采用人工清点载矿矿车数的方式估算出矿量,该计量方式存在估算误差大等缺点,严重影响了矿山企业生产计划的合理制定。本文采用图像识别与密度模型建模相结合的方法,通过对矿石堆图像信息进行三维重建形成体积模型,构建了一个图像特征—密度库,设计并开发了一套基于图像识别的井下机车载矿量计量系统。经现场反复试验证明,该计量系统运行稳定可靠,机车载矿量计算误差小于5%,较好地解决了目前矿山企业出矿量估算不准确的问题,提高了出矿量计算的准确度,为矿山企业制定生产计划提供了详实可靠的数据。

关键词: 图像识别, 三维重建, 密度模型, 计量系统, 井下机车, 出矿量

Abstract:

The ore yield is an important standard for mining enterprises to formulate production plans.At present,the vast majority of underground mines estimate ore yield by manually counting the number of ore carrying vehicles.The measurement error is large,which seriously affects the enterprises to formulate production plans.In order to solve the problem of large measurement error and improve the accuracy of ore yield estimation,a set of underground machine on-board ore yield measurement system based on image recognition was designed and developed in this paper.In this paper,the method of image recognition combined with density model modeling was used to form a volume model through three-dimensional reconstruction of ore pile image information,and an image feature density library was built to form a complete set of underground machine on-board ore quantity measurement system.The system collects the internal image of the locomotive ore bucket through the depth camera,then extracts the feature information of the image and compares it with the image feature density library to obtain the density of the ore in the current ore bucket, and then generates a volume model from the three-dimensional reconstruction of the image to calculate the volume of the ore pile,and calculates the product of the volume and density of the ore pile to obtain the weight of the ore pile.The field repeated tests show that the metering system operates stably and reliably,and the calculation error of locomotive ore load is less than 5%.It solves the problem of ore yield estimation in mining enterprises,improves the calculation accuracy of ore yield,and brings detailed and reliable data for enterprises to formulate production plans.

Key words: image recognition, 3D reconstruction, density model, measurement system, underground locomotive, ore yield

中图分类号: 

  • TP274+.5

图1

井下机车载矿量计量系统结构图"

图2

相机集成装置1-补偿灯;2-灯架;3-灯板;4-透明板;5-支撑板;6-开关电源;7-深度相机;8-光纤收发器;9-箱体;10-支架;11-电源接口;12-网线接口;13-光纤转换接口"

图3

矿车与深度相机的安装示意图"

图4

深度图定义"

图5

矿车图像效果图"

图6

载矿量实时监测系统界面"

表1

部分矿车总重量、矿石体积及平均密度"

矿车总重量/kg矿石重量/kg矿石体积/m3

平均密度

/(×103 kg·m-3

1 7601 0320.8411.227
1 7951 0670.8711.224
1 7951 0670.8681.229
1 8661 1380.5312.142
1 8661 1380.8111.404
1 9731 2450.6202.009
1 9731 2450.6981.783
1 9731 2450.6561.898
1 9731 2450.7441.673
1 8901 1620.6401.814
1 8901 1620.6531.780
1 9701 2420.7381.683
1 9701 2420.7571.640
1 9891 2610.4752.653
2 1121 3840.8511.626
2 0421 3140.8161.611

图7

矿车内矿石堆的平均密度折线图"

表2

部分图像文件名与平均密度对应关系"

图像文件名平均密度/(×103 kg·m-3
2020-11-13-09-37-58-887_RGB.png1.227
2020-11-13-09-38-02-599_RGB.png1.224
2020-11-13-09-38-02-789_RGB.png1.229
2020-11-13-09-38-33-286_RGB.png2.142
2020-11-13-09-38-36-077_RGB.png1.404
2020-11-13-09-38-41-159_RGB.png2.009
2020-11-13-09-38-42-699_RGB.png1.784
2020-11-13-09-38-45-640_RGB.png1.898
2020-11-13-09-38-46-571_RGB.png1.674
2020-11-13-09-38-53-163_RGB.png1.815
2020-11-13-09-38-58-121_RGB.png1.781
2020-11-13-09-38-59-648_RGB.png1.684
2020-11-13-09-39-08-730_RGB.png1.640
2020-11-13-09-53-39-621_RGB.png2.653
2020-11-13-09-53-50-114_RGB.png1.626
2020-11-13-09-54-03-006_RGB.png1.611

表3

部分计量系统测量结果与轨道衡称重结果对比"

矿石体积

/m3

计量系统测量结果

/kg

轨道衡称重结果

/kg

相对误差

/%

0.7872011 2151 1446.21
0.7269661 1671 0986.28
0.4245921 0221 088-6.07
0.7690641 1011 0386.07
0.7493071 1601 0995.55
0.7421111 1611 0975.83
0.7872051 0941 0326.01
0.7927871 2301 1675.40
0.4753871 1591 229-5.70
0.8513791 5061 4215.98
0.8155181 2591 314-4.19
0.7977621 1141 182-5.75
0.8485251 4041 3305.56
0.8168211 3851 3314.06
0.8462121 4391 3566.12
0.8631801 4141 3484.90
Baklanova O E, Baklanov M A,2016.Methods and algorithms of image recognition for mineral rocks in the mining industry[J].Advances in Swarm Intelligence,9713:253-262.
Chen Shaojie, Li Guangli, Zhang Wei, al et,2011.Land use classification in coal mining area using remote sensing images based on multiple classifier combination[J].Journal of China University of Mining & Technology,40(2):273-278.
Dai Changlu,2020.Comprehensive application of XRT ray and image intelligent concentrator in wolframite mine[J].China Metal Bulletin,(11):187-188.
Di Wangping, Wu Zhihu,2021.Preconcentration and discarding technology of intelligent photoelectric dressing equipment[J].Nonferrous Metals:Mieral Processing Section,(1):117-121.
Du Peijun, Liu Sicong, Zheng Hui,2012.Land cover change detection over mining areas based on support vector machine[J].Journal of China University of Mining & Technology,41(2):262-267.
Guo Yi, Wu Fuji, Zhong Yi, al et,2021.Research on intelligent cleaning device for bottom of tipping-type car based on image recognition technology[J].China Tungsten Industry,36(2):76-80.
He Jie, Wang Guimei, Liu Jiehui, al et,2020.Volume measurement of coal volume on belt conveyor based on image processing[J].Acta Metrologica Sinica,41(12):1516-1520.
Li Guoqing, Li Bao, Hu Nailian, al et,2017.Optimization model of mining operation scheduling for underground metal mines[J].Chinese Journal of Engineering,39(3):342-348.
Li Hexian,2020.Microscopic measurement and refreshing device for CCD camera[J].Journal of Chongqing University of Science and Technology(Natural Sciences Edition),22(2):93-96.
Li Xingdong, Zhang Rui,2021.On differential element method and its principle[J].Studies in College Mathematics,24(1):80-83.
Liang Le,2019.Research on Measurement Method of Irregular Object Volume Based on Binocular Stereo Vision[D].Xi’an:Xi’an University of Technology.
Ling Xiaoming, Guo Ruixin, Liu Guangting, al et,2021.Pedestrian detection research based on multi-feature cascade of PCA dimension reduction[J].Manufacturing Automation,43(3):32-34,76.
Patel A K, Chatterjee S, Gorai A K,2019.Development of a machine vision system using the support vector machine regression (SVR) algorithm for the online prediction of iron ore grades[J].Earth Science Informatics,12(2):197-210.
Ren Zhiwei, Wu Linda,2018.Feature Extraction for hyperspectral images based on principal component analysis improved by information quantity[J].Journal of Sichuan Ordnance,39(7):151-154.
Sun X Y, Li X R, Xiao D, al et,2021.A method of mining truck loading volume detection based on deep learning and image recognition[J].Sensors,21(2):635.
Tian Tian,2011.Coal Petrography Analysis and Coal Pile Volume Measurement Based on Image Processing Technology[D].Taiyuan:Taiyuan University of Technology.
Wang Fang,2020.Research on Primary Selection System of Black Tungsten Ore Based on Machine Vision[D].Ganzhou:Jiangxi University of Science and Technology.
Wang Jun, Pu Lei, He Xinyu, al et,2021.Multiple biological features recognition algorithms of underground coalmine sign-in system[J].Bulletin of Science and Technology,37(3):44-49.
Wang Liguan, Chen Sijia, Jia Mingtao, al et,2020.Beneficiation method of wolframite image recognition based on deep learning[J].The Chinese Journal of Nonferrous Metals,30(5):1192-1201.
Wang Yaoge, Zhang Dongyan, Zhang Ning,2020.Differential element method for definite integrals[J].Studies in College Mathematics,23(6):7-9,20.
Wang Yixin,2016.Application of depth-image processing in vehicle identification[J].Value Engineering,35(25):236-238.
Xiao Jiwei,2019.Research and Design of Black Tungsten Ore Intelligent Sorting System Based on Machine Vision[D].Changsha:Hunan University.
Xu J K, Wang E Y, Zhou R,2020.Real-time measuring and warning of surrounding rock dynamic deformation and failure in deep roadway based on machine vision method[J].Measurement,149:107028.
Yang Boxiong, Yang Yuqi,2019.Applying PCA to dimensionality reduction of image features extracted by deep learning[J].Computer Systems & Applications,28(1):279-283.
Yang Jintao,2019.Research and development of mine electric locomotive[J].World Nonferrous Metals,4(4):234,236.
Yang Wenlong, Ma Baoliang, Chen Chen,2019.A method of measuring the ore carrying capacity of hopper based on depth camera[J].China Tungsten Industry,34(6):69-74.
Ye Zhenglin, Lin Wei,2020.Application of the differential element method in some integral problems[J].Studies in College Mathematics,23(4):32-37.
Zhang Aihua, Tang Tingting, Wang Weiwei, al et,2018.A fast fractal image compression algorithm based on principal component[J]. Computer Technology and Development,28(5):77-80,85.
Zhang Hao, Zhang Qiang, Li Yongxiang, al et,2021.Research on 3D model reconstruction based on deep learning[J].Jou-rnal of Chongqing University of Posts and Telecommunications(Natural Science Edition),33(2):289-295.
Zhang Jian,2017.Application of the differential element method in the volume calculation for the rotating bodies[J].College Mathematics,33(4):104-110.
Zhang Jianli, Ye Pingkun, Sun Shenshen,2020.Detection of ore granularity based on morphological image processing[J].Machinery Design and Manufacture,4(3):68-71.
Zhang Xiaoniu,2016.Design of cleaning car bottom car hydraulic control system in screw type[J].Development & Innovation of Machinery & Electrical Products,29(4):102-103,22.
Zhang Zelin, Yang Jianguo, Su Xiaolan, al et,2013.Estimation of density distribution of coarse coal pile by image analysis[J].Journal of China University of Mining & Technology,42(5):851-858.
Zheng K H, Du C L, Li J P, al et,2015.Underground pneumatic separation of coal and gangue with large size(≥50 mm) in green mining based on the machine vision system[J].Powder Technology,278:223-233.
陈绍杰,李光丽,张伟,等,2011.基于多分类器集成的煤矿区土地利用遥感分类[J].中国矿业大学学报,40(2):273-278.
戴昌璐,2020.XRT射线及图像智能选矿机在黑钨矿山的综合运用探索[J].中国金属通报,(11):187-188.
第旺平,吴志虎,2021.智能光电选矿预选抛废技术研究及应用[J].有色金属(选矿部分),(1):117-121.
杜培军,柳思聪,郑辉,2012.基于支持向量机的矿区土地覆盖变化检测[J].中国矿业大学学报,41(2):262-267.
郭毅,吴富姬,钟毅,等,2021.基于图像识别技术的翻斗式矿车结底智能清理装置研究[J].中国钨业,36(2):76-80.
贺杰,王桂梅,刘杰辉,等,2020.基于图像处理的皮带机上煤量体积计量[J].计量学报,41(12):1516-1520.
李国清,李宝,胡乃联,等,2017.地下金属矿山采掘作业计划优化模型[J].工程科学学报,39(3):342-348.
李和仙,2020.CCD相机的微距测量补光装置设计[J].重庆科技学院学报(自然科学版),22(2):93-96.
李兴东,张睿,2021.关于微元法及其原理的探讨[J].高等数学研究,24(1):80-83.
梁乐,2019.基于双目立体视觉的不规则物体体积测量方法研究[D].西安:西安理工大学.
令晓明,郭锐辛,刘光廷,等,2021.基于PCA降维的多特征级联的行人检测研究[J].制造业自动化,43(3):32-34,76.
任智伟,吴玲达,2018.基于信息量改进主成分分析的高光谱图像特征提取方法[J].兵器装备工程学报,39(7):151-154.
田甜,2011.基于图像处理技术的煤岩分析及煤堆体积测量[D].太原:太原理工大学.
汪钇鑫,2016.深度图像处理在车辆识别中的应用[J].价值工程,35(25):236-238.
王芳,2020.基于机器视觉的黑钨矿石初选系统研究[D].赣州:江西理工大学.
王君,蒲磊,何新宇,等,2021.多生物特征融合的矿井人员身份识别[J].科技通报,37(3):44-49.
王李管,陈斯佳,贾明滔,等,2020.基于深度学习的黑钨矿图像识别选矿方法[J].中国有色金属学报,30(5):1192-1201.
王耀革,张冬燕,张宁,2020.微元分析法探究[J].高等数学研究,23(6):7-9,20.
肖继伟,2019.基于机器视觉的黑钨矿石智能分选系统研究与设计[D].长沙:湖南大学.
杨博雄,杨雨绮,2019.利用PCA进行深度学习图像特征提取后的降维研究[J].计算机系统应用,28(1):279-283.
杨锦涛,2019.矿山工矿电机车研究与发展[J].世界有色金属,4(4):234,236.
杨文龙,马保亮,陈辰,2019.基于深度相机的矿斗载矿量的测量方法[J].中国钨业,34(6):69-74.
叶正麟,林伟,2020.微元法在一些积分问题中的应用[J].高等数学研究,23(4):32-37.
张爱华,唐婷婷,汪玮玮,等,2018.基于主成分特征的快速分形图像压缩算法[J].计算机技术与发展,28(5):77-80,85.
张豪,张强,李勇祥,等,2021.基于深度学习的三维模型重构研究[J].重庆邮电大学学报(自然科学版),33(2):289-295.
张建立,叶平坤,孙深深,2020.形态学图像处理下的矿石粒度的检测[J].机械设计与制造,4(3):68-71.
张健,2017.旋转体体积计算中的微元法思想应用[J].大学数学,33(4):104-110.
张小牛,2016.螺旋式清理矿车结底车液压控制系统的设计[J].机电产品开发与创新,29(4):102-103,22.
张泽琳,杨建国,苏晓兰,等,2013.基于图像分析的粗粒煤堆密度组成估计[J].中国矿业大学学报,42(5):851-858.
No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
[1] 肖风利,曾庆栋,马凤山,王昭坤,孙之夫,孙宗峰. 胶东西北部重要金成矿断裂带特征[J]. 黄金科学技术, 2018, 26(4): 396 -405 .
[2] 金鹏,刘科伟,李旭东,杨家彩. 深部岩体水耦合爆破裂纹扩展数值模拟研究[J]. 黄金科学技术, 2021, 29(1): 108 -119 .
[3] 王久玲,李国清,胡乃联,陈博宇. 基于熵权—CTODIM模型的“三下”矿体采矿方法优选[J]. 黄金科学技术, 2016, 24(4): 81 -86 .
[4] 江思宏, 张莉莉, 刘翼飞, 李高峰, 季根源. 非洲大陆金矿分布特征与勘查建议[J]. 黄金科学技术, 2020, 28(4): 465 -478 .
[5] 王宇慧,罗先熔,何旺,王东,汤国栋,商振城. 广西罗城县记洞湾矿区地电化学提取法寻找隐伏锡铜多金属矿的研究及找矿预测[J]. 黄金科学技术, 2021, 29(4): 500 -509 .
[6] 刘永春,王李管,吴家希. 基于LQR-QPSO的地下铲运机控制参数优化研究[J]. 黄金科学技术, 2021, 29(1): 25 -34 .
[7] 范红科,郭根明,董波波,张凯,刘豫新. 新疆焉耆县松树达坂金矿区岩石地球化学异常特征及找矿效果[J]. 黄金科学技术, 2021, 29(4): 477 -488 .
[8] 于世波, 杨小聪, 原野, 王志修. 深部区域采矿时序的地压调控卸荷效应研究[J]. 黄金科学技术, 2020, 28(3): 345 -352 .
[9] 毕林,王黎明,段长铭. 矿井环境高精定位技术研究现状与发展[J]. 黄金科学技术, 2021, 29(1): 3 -13 .
[10] 景岳,王少锋,鲁金涛. 矿岩开挖松动区厚度预测及非爆机械化开采判据[J]. 黄金科学技术, 2021, 29(4): 525 -534 .