煤矸分拣机器人在工作时，由于带式输送机出现打滑、跑偏等现象，导致煤矸石定位不准确而造成机器人抓取失败，甚至由于冲击载荷造成机械臂损坏。针对该问题，本文提出了基于视觉伺服的机械臂动态目标跟踪方法。

       针对动态煤矸石精准定位问题，构建了“机器人-机械臂末端-相机-矸石”定位模型，获得了目标矸石在机器人全局坐标系下的位姿信息，实现了动态目标矸石的精准定位。根据煤矸分拣机器人本体特点，建立了煤矸分拣机器人数学模型，根据模型进行运动学逆问题求解，解算所需要的机器人机身位姿信息，作为完成机械臂精确控制的运动基础。

       针对视觉跟踪目标矸石问题，提出一种基于Hu不变矩的动态目标快速匹配方法来快速获取目标矸石的位置信息，实现矸石信息从识别模块到视觉伺服模块的传递，得到目标矸石初步位置信息。针对视觉在对目标矸石进行跟踪过程中因遮挡而导致跟踪丢失问题，提出一种结合卡尔曼滤波的扩展FDSST（快速判别尺寸空间跟踪算法）视觉跟踪定位方法，实现目标矸石实时位置信息的快速获取。

       针对煤矸分拣机械臂同步跟踪动态目标矸石问题，提出一种基于视觉伺服的三环PID控制机械臂动态目标跟踪轨迹规划方法。该方法以视觉获取的目标矸石实时精确位置作为位置环控制器的输入，位置环控制器的输出作为速度环控制器的输入，速度环控制器的输出作为加速度环控制器的输入，实现位置、速度和加速度的最优控制，使得机械臂末端和目标矸石到达同步，最终实现机械手拟静态抓取矸石，提高了抓取成功率。

       在上述提出的煤矸分拣机器人精确定位以及同步跟踪方法的基础上，在实验室搭建的煤矸分拣机器人实验平台上进行实验验证。结果表明：本文所提出的基于视觉伺服的机械臂同步跟踪煤矸石方法，在不同带速下均能对目标矸石进行有效跟踪，视觉跟踪成功率可达到97%以上。三环PID控制轨迹规划算法平均位置误差较小可达到1 mm左右，在抓取点的平均速度误差在1 mm/s左右，跟踪速度误差较小，可满足对高速度目标的同步跟踪、精准抓取要求。

关 键 词：煤矸分拣机械臂；视觉伺服；动态模板匹配；扩展FDSST跟踪算法；三环PID控制轨迹规划

        When the coal gangue sorting robot is working, due to the slippage and deviation of the belt conveyor, the positioning of the coal gangue is inaccurate and the robot fails to grasp, and even the mechanical arm is damaged due to the impact load. Aiming at this problem, this paper proposes a dynamic target tracking method for robotic arms based on visual servoing.

        Aiming at the problem of precise positioning of dynamic coal gangue, a "robot-manipulator end-camera-gangue" positioning model was constructed, and the position and attitude information of the target gangue in the robot coordinate system was obtained, and the precise positioning of the dynamic target gangue was realized. According to the characteristics of the coal gangue sorting robot body, a mathematical model of the coal gangue sorting robot is established, and the inverse kinematics problem is solved according to the model, and the required pose information of the robot body is calculated as the motion basis for the precise control of the manipulator.

        Aiming at the problem of visual tracking target gangue, a dynamic target fast matching method based on Hu invariant moment is proposed to quickly obtain the position information of the target gangue, realize the transfer of the gangue information from the recognition module to the visual servo module, and obtain the initial position information of the target gangue. Aiming at the problem of tracking loss caused by occlusion in the process of tracking the target gangue, an extended FDSST (fast discriminative size space tracking algorithm) visual tracking and positioning method combined with Kalman filter is proposed to realize the rapid acquisition of the real-time position information of the target gangue.

        Aiming at the problem of gangue synchronously tracking dynamic target gangue for coal gangue sorting manipulator, this paper proposes a dynamic target tracking trajectory planning method based on visual servo three-loop PID control manipulator. The method uses the real-time precise position of the target gangue obtained visually as the input of the position loop controller, the output of the position loop controller is used as the input of the velocity loop controller, and the output of the velocity loop controller is used as the input of the acceleration loop controller, so as to realize the position, The optimal control of speed and acceleration makes the end of the manipulator and the target gangue reach synchronization, and finally realizes the quasi-static grasping of the gangue by the manipulator, which improves the success rate of grasping.

        On the basis of the precise positioning and synchronous tracking method of the coal gangue sorting robot proposed above, the experimental verification is carried out on the experimental platform of the coal gangue sorting robot built in the laboratory. The results show that the proposed method for synchronously tracking coal gangue based on visual servoing can effectively track the target gangue under different belt speeds, and the success rate of visual tracking can reach more than 97%. The average position error of the three-loop PID control trajectory planning algorithm is as small as about 1 mm, the average speed error at the grab point is about 1 mm/s, and the tracking speed error is small, which can meet the requirements of synchronous tracking and accurate tracking of high-speed targets. Crawl request.

Keywords:Coal gangue sorting robotic arm; Visual servoing; Dynamic template matching; Extended FDSST tracking algorithm; Three-loop PID control trajectory planning

[1]康利,黄金辉,刘畅.GDRT型γ射线智能干法分选系统在六家煤矿的应用[J].煤炭加工与综合利用,2017(3):22-24.

[2]郭铁成.中国制造2025:智能时代的国家战略[J].人民论坛·学术前沿,2015(19):54-67.

[3]尹建平.平朔矿区煤矸石山生态修复模式[J].露天采矿技术,2021,36(04):87-88+92.

[4]陈恳.基于机器视觉的物料分拣工业机器人关键技术研究[D].深圳:深圳大学,2017.

[5]郭宾宾,付令怡,庞明瑾. 选煤厂矸石外排清洁生产技术研究与应用[J]. 洁净煤技术, 2013, 19(05):103-104+119.

[6]Wang XP, Shang B, Yuan ZJ, et al. Discussion on how to improve the quality of raw coal before washing[J]. Coal Quality Technology, 2014.

[7]王国法,刘峰,庞义辉,等.煤矿智能化煤炭工业高质量发展的核心技术支撑[J].煤炭学报,2019,44(02):349-357.

[8]辛绍磊,马文伟.干法选煤技术在选煤厂的应用及局限性[J].煤炭加工与综合利用,2020(09):17-19+23.

[9]靳璞磊.基于智能相机和工业机器人的自动分拣技术研究[D]. 南京: 南京航空航天大学, 2016.

[10]贾贺林.基于DTW的人体行为模式识别算法[J].电子技术与软件工程,2019(05):61.

[11]Burger W, Burge M J. Digital image processing[M]. Springer, 2008.

[12]邹广华.基于几何特征的快速模板匹配算法[D].哈尔滨工业大学,2010.

[13]王刚,孙晓亮,尚洋,等.一种基于最佳相似点对的稳健模板匹配算法[J].光学学报,2017,37(03):281-287.

[14]王凤梅,卢文彪,陈仕妍.基于灰度匹配模板的中药材显微图像识别[J].中国实验方剂学杂志,2019,25(11):167-172.

[15]Zhou Fang et al.SAR Target Detection Based on Improved SSD with Saliency Map and Residual Network[J]. Remote Sensing, 2022, 14(1) : 180-180.

[16]贾迪,朱宁丹,杨宁华,等.图像匹配方法研究综述[J].中国图象图形学报,2019,24(05):677-699.

[17]倪佳忠,马志龙.局部特征信息的航拍图像匹配算法[J].北京绘,2022,36(03):228-232.

[18]高达义.基于局部特征的人脸表情识别算法[D].南京邮电大学,2021.

[19]De Castro E,Morandi C. Registration of translated and rotated images using finite fourier transforms.[J]. IEEE transactions on pattern analysis and machine intelligence,1987,9(5).

[20]王雅雯.轮毂型号在线识别系统研究[D].长春理工大学,2020.

[21]MadHuri Yadav,Ravindra Kumar Purwar. Hindi handwritten character recognition using oriented gradients and Hu-geometric moments[J]. Journal of Electronic Imaging,2018,27(5).

[22]Rachid Benouini,Imad Batioua,Khalid Zenkouar,et al. Fast and accurate computation of Racah moment invariants for image classification[J]. Pattern Recognition,2019,91.

[23]李鑫,赵坤鹏,朱凌寒.改进的不变矩和PNN相结合的多品种产品识别算法[J].传感器与微系统,2019,38(08):132-135.

[24]Sardar Jehangir,Sohail Khan,Sulaiman Khan, et al. Zernike Moments Based Handwritten Pashto Character Recognition Using Linear Discriminant Analysis[J]. Mehran University Research Journal of Engineering and Technology,2021,40(1).

[25]黄浩然.基于Hu不变矩的垃圾分类和识别[J].自动化应用,2020(08):74-76.

[26]伍蔚帆.基于卷积神经网络的目标跟踪方法研究[D].重庆邮电大学,2020.DOI:10.27675/d.cnki.gcydx.2020.000461.

[27]彭建盛,许恒铭,李涛涛,等.生成式与判别式视觉目标跟踪算法综述[J].科学技术与工程,2021,21(35):14871-14881.

[28]Guo Xifeng,Hamdulla Askar,Tohti Turdi. Research on Target Tracking Algorithm Based on Correlation Filtering[J]. Journal of Physics: Conference Series,2021,2024(1).

[29]孙金萍.复杂场景下视频目标跟踪算法研究[D].中国矿业大学,2021.DOI:10.27623/d.cnki.gzkyu.2021.000016.

[30]杨亚男,付春玲.视频目标跟踪算法综述[J].科技资讯,2018,16(02):14-15.

[31]朱文青,刘艳,卞乐,等.基于生成式模型的目标跟踪方法综述[J].微处理机,2017,38(01):41-47.

[32]周丽君.基于深度学习的目标跟踪算法研究[D].中国科学院大学(中国科学院光电技术研究所),2021.

[33]杨宇翔.基于深度强化学习的目标跟踪算法研究[D].北京交通大学,2021.

[34]Pei Lili,Zhang He,Yang Bo. Improved Camshift object tracking algorithm in occluded scenes based on AKAZE and Kalman.[J]. Multimedia tools and applications,2021,81(2).

[35]Bolme D S，Beveridge J Ｒ，Draper B A，et al． Visual object tracking using adaptive correlation lters［C］/ /Proceedings of the IEEE Conference on Computer Vision and Pa.

[36]单明媚,王华通,郑浩岚,等.视频运动目标跟踪算法研究综述[J].物联网技术,2019,9(07):76-78+82.

[37]徐宁,王娟娟,郭晓雨,等.判别式相关滤波器的目标跟踪综述[J].小型微型计算机系统,2020,41(12):2484-2493.

[38]陈彦全.基于判别式模型的目标跟踪方法研究[D].重庆邮电大学,2018.DOI:10.27675/d.cnki.gcydx.2018.000312.

[39]Danelljan M，Hager G，Khan F，et al． Accurate scale estimation for robust visual tracking［C］/ /British Machine Vision Confer- ence． Nottingham: BMVA Press，2014: 1-11.

[40]Danelljan M，Hager G，Khan F S，et al． Discriminative scale space tracking［J］． IEEE Transactions on Pattern Analysis and Machine Intelligence，2017，39( 8) : 1561-1575.

[41]Farhodov, Xurshid, Kwon,et al.A New CSR-DCF Tracking Algorithm based on Faster RCNN Detection Model and CSRT Tracker for Drone Data[J].JOURNAL OF KOREA MULTIMEDIA SOCIETY, 22 (12) , pp.1415-1429.

[42]王彩霞,林寿英.基于深度学习的多目标跟踪算法综述[J].中阿科技论坛(中英文),2021(10):118-120.

[43]Dong-mei Shi,Xun Chen. Research on visual object tracking algorithm based on improved twin network[J]. Journal of Physics: Conference Series,2021,1966(1).

[44]王成军,严晨.机器视觉技术在分拣系统中的应用研究综述[J].制造技术与机床,2020(05):32-37.

[45]Liu, Yiwen,Zhang,et al. Intelligent robot motion trajectory planning based on machine vision[J]. International Journal of System Assurance Engineering and Management,2022(prepublish).

[46]Rana A，Valenzise G，Dufaux F．Learning-based tone mapping operator for image matching［J］．IEEE Transactions on Multimedia，2018(99):256-268．

[47]董理,杨东,鹿建森.工业机器人轨迹规划方法综述[J/OL].控制工程:1-12[2022-04-27].

[48]葛利刚.基于视觉的快速运动目标跟踪及抓取问题研究[D].国防科学技术大学,2016.

[49]苏剑波.机器人抓取运动目标轨迹规划方法[J].机器人,1994,16(2):71-76.

[50]张朝阳,周惠兴,曹荣敏,等.基于金字塔形寻优的传送带动态抓取研究[J].制造业自动化,2015,37(09):92-95.

[51]曹现刚,李宁,王鹏,等.基于比例导引法的机械臂拣矸过程轨迹规划方法研究[J].煤炭工程,2019,51(05):154-158.

[52]袁晗,童上高,顾文昊,等.机器人动态物料追随抓取技术的轨迹算法设计[J].机器人技术与应用,2019(01):29-32.

[53]李宁.煤矸分拣机器人控制系统研究[D].西安科技大学,2019.

[54]王鹏,曹现刚,马宏伟,等. 基于余弦定理-PID的煤矸石分拣机器人动态目标稳准抓取算法[J]. 煤炭学报,2020(12):4240-4247.

[55]李雪锋,庄健培,高燕,等.一种机器人的传送带动态跟踪算法[J]. 机床与液压,2021,49(17):65-67,73. DOI:10.3969/j.issn.1001-3881.2021.17.011.

[56]Tao Y, Ren F, Gao H, et al. A Robot Dynamic Target Grasping Method Based on Affine Group Improved Gaussian Resampling Particle Filter. Applied Sciences.2021.

[57]Yi-CHun Du,Taryudi Taryudi,Ching-Tang Tsai,et al. Eye-to-hand robotic tracking and grabbing based on binocular vision[J]. Microsystem Technologies,2019,27(prepublish).

[58]XiaoHui Xie, CHunyu ZHu, Ming Xie. Visual tracking control of SCARA robot system based on deep learning and Kalman prediction method[J]. International Journal of Hydromechatronics,2021,4(4).

[59]卢艳军,王诗宇,张太宁,等.针对目标跟踪过程的双目视觉测距方法研究[J].电光与控制,2020,27(1):60-63+101.

[60]计算机视觉[M]. 科学出版社 , 马颂德,张正友著, 1998.

[61]姜雨濛.基于Kinect的工件识别方法研究[D].沈阳工业大学,2020.

[62]姚阳.煤矿巷道龙门式钻锚机器人精确定位技术研究[D].西安科技大学,2021.

[63]马梦婕.基于双目立体视觉的动态物料辨识方法的研究[D].西南交通大学,2018.

[64]李金灿.基于视觉反馈的机器人控制算法研究[D].杭州电子科技大学,2019.

[65]苗继松,邵琼玲,任元.基于三元角的坐标旋转变换方法[J].北京航空航天大学学报,2017,43(12):2539-2546.

[66]房国栋,高军伟,朱晨曦,等.基于机器视觉的机械臂智能分拣系统[J].仪表技术与传感器,2020(12):72-76.

[67]Cai Jie,Deng Jinlian,Zhang Wei, et al. Modeling Method of Autonomous Robot Manipulator Based on D-H Algorithm[J]. MOBILE INFORMATION SYSTEMS,2021,2021.

[68]李莹.基于深度学习的煤矸石目标检测方法研究[D].西安科技大学,2020.

[69]黄浩然.基于Hu不变矩的垃圾分类和识别[J].自动化应用,2020(8):74-76.

[70]Hicham Tribak,Youssef Zaz. QR Code Patterns Localization based on Hu Invariant Moments[J]. International Journal of Advanced Computer Science and Applications (IJACSA),2017,8.

[71]D. V. Kondusov,A. I. Sergeev,V. B. Kondusova. Comparison of 3D Models Using Hu Moment Invariants[J]. Russian Engineering Research,2020,40(7).

[72]顾梅花,苏彬彬,王苗苗,等.彩色图像灰度化算法综述[J].计算机应用研究,2019,36(05):1286-1292.DOI:10.19734/j.issn.1001-3695.2018.01.0125.

[73]刘利平,乔乐乐,蒋柳成.图像去噪方法概述[J].计算机科学与探索,2021,15(08):1418-1431.

[74]石炜,张袁祥.轴承缺陷图像二值化处理及形态学边缘检测[J].舰船电子工程,2020,40(08):114-117.

[75]钱月.基于MATLAB的图像形态学处理技术与应用[J].内江师范学院学报,2019,34(10):51-55.

[76]白蕾,王聪慧,蒋超.基于早期认知视觉的图像轮廓提取[J].国外电子测量技术,2019,38(12):16-20.

[77]黄超,茅健,徐斌,周玉凤.基于最小外接矩形和Hough变换的定位算法[J].组合机床与自动化加工技术,2021(08):66-71.

[78]余铎,王耀南,毛建旭,等.基于视觉的移动机器人目标跟踪方法[J].仪器仪表学报,2019,40(01):227-235.

[79]Liang ZHujun,Zhang Xingping,Shao YanHua. An Improved Anti-occlusion FDSST Algorithm with SSDA[J]. TWELFTH INTERNATIONAL CONFERENCE ON GRAPHICS AND IMAGE PROCESSING (ICGIP 2020),2021,11720.

[80]李轶锟,吴庆宪,丁晟辉,等.基于TLD和fDSST的长时间目标跟踪算法[J].电光与控制,2019,26(04):44-48+70.

[81]WANG M，SU D，SHI L，et al． Real － time 3D Human tracking for mobile robots with multisensors［C］． IEEE International Conference on Robotics and Automation ( ICRA) ，2017: 5081-5087．

[82]代天乐.复杂遮挡环境下目标检测与跟踪技术研究[D].长春理工大学,2021.DOI:10.26977/d.cnki.gccgc.2021.000341.

[83]吕德深,梁承权.模糊理论和PID相结合的激光雷达智能控制系统[J].激光杂志,2021,42(07):157-160.

[84]徐涢基,龚明,龚诚,等.基于双环PID算法与大津法的板球视觉控制系统[J].传感器与微系统,2020,39(11):81-84.

[85]王晓海,孟秀云,周峰,等.基于偏置比例导引的落角约束滑模制导律[J].系统工程与电子技术,2021,43(05):1295-1302.