移动机器人视觉跟踪技术的研究

发布时间：2018-03-25 05:29

本文选题：移动机器人　切入点：单目视觉测距　出处：《吉林大学》2017年硕士论文

【摘要】：近些年,随着国内外机器人产业飞速发展以及人工智能领域取得巨大进步,机器人在人类的生活中扮演的角色越来越重要。与此同时近些年国家提出建立万物感知、万物互联、万物智能的智慧城市和平安城市,使国内安全防范行业获得高速发展的机遇,从而使机器视觉研究持续火热,机器人可以通过视觉系统来获取大量信息,继而通过对信息的处理、加工,最后做出决策。本文是以移动机器人作为研究对象,主要研究如何通过视觉跟踪算法,实现跟踪机器人对领航机器人在各种运动模式下,均能够实现准确且稳定的跟踪。通过视觉跟踪控制系统以及机器人运动控制系统,来实现移动机器人的跟踪任务。本文通过阅读大量文献和实验主要完成了如下三个任务:(1)详细介绍了机器人视觉跟踪控制系统的研究意义、背景及国内外研究现状,然后对单目视觉系统的基本原理进行了深入研究,其中包括单目摄像机的标定技术,以及单目视觉测距原理等,最后通过实验验证了单目视觉测距原理的准确性。(2)摄像机视觉跟踪算法的研究。首先对均值漂移算法(Mean Shift)和自适应均值漂移算法(CamShift)进行研究分析,然后将粒子滤波算法运用到跟踪领域中,通过对比实验来分析CamShift跟踪算法和粒子滤波算法的优缺点,最后提出将粒子滤波算法和CamShift算法相结合的算法作为本文的跟踪算法,通过实验证明该算法能够提高视觉目标跟踪的准确性和鲁棒性,为移动机器人视觉跟踪控制系统提供稳定的视觉跟踪算法。(3)移动机器人视觉跟踪控制系统的实现。首先通过视觉跟踪算法,获取跟踪目标的信息,将目标返回的信息作为移动机器人运动的依据,以此来设计视觉跟踪的控制策略,然后将跟踪机器人与领航机器人之间的距离信息和领航机器人的质心变化作为判断的依据,最后设计运动控制策略,使得领航机器人始终在跟踪机器人的视野中,从而实现移动机器人视觉跟踪控制系统。最后,本文基于OpenCV图像处理库和Microsoft Visual C++6.0进行软件开发,实现了跟踪机器人对领航机器人在实验室环境中的有效稳定的跟踪。
[Abstract]:In recent years, with the rapid development of robot industry at home and abroad and the great progress in artificial intelligence field, robot plays an increasingly important role in human life. The intelligent city of all things and the safe city make the domestic security guard industry gain the opportunity of high speed development, so that the machine vision research continues to be hot, the robot can obtain a large amount of information through the vision system. Then through the information processing, processing, and finally making a decision. This paper takes the mobile robot as the research object, mainly studies how to realize the tracking robot to the pilot robot in various motion modes through the visual tracking algorithm. Through visual tracking control system and robot motion control system, By reading a large number of literature and experiments, this paper mainly completed the following three tasks: 1) introduced the research significance, background and research status of robot visual tracking control system in detail. Then the basic principle of monocular vision system is deeply studied, including the calibration technology of monocular camera and the principle of monocular vision ranging. Finally, the accuracy of the principle of monocular vision ranging is verified by experiments. Firstly, the mean shift algorithm (mean shift) and the adaptive mean shift algorithm (CamShift) are studied and analyzed. Then the particle filter algorithm is applied to the field of tracking, and the advantages and disadvantages of CamShift tracking algorithm and particle filter algorithm are analyzed through comparative experiments. Finally, the algorithm combining particle filter algorithm and CamShift algorithm is proposed as the tracking algorithm in this paper. Experiments show that the algorithm can improve the accuracy and robustness of visual target tracking. This paper provides a stable visual tracking algorithm for mobile robot visual tracking control system. Firstly, the visual tracking control system of mobile robot is realized. Firstly, the information of tracking target is obtained by visual tracking algorithm. The information returned by the target is taken as the basis for the mobile robot to design the control strategy of visual tracking. Then the distance information between the tracking robot and the pilot robot and the change of the centroid of the pilot robot are taken as the basis of judging. Finally, the motion control strategy is designed so that the pilot robot is always in the field of vision of the tracking robot, so as to realize the visual tracking control system of the mobile robot. Finally, this paper develops the software based on OpenCV image processing library and Microsoft Visual C 6.0. The tracking robot can track the pilot robot effectively and stably in the laboratory environment.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TP242

【参考文献】