【作者】 王乐；

【导师】 焦恩璋；

【作者基本信息】 南京林业大学 ， 控制理论与控制工程， 2012， 硕士

【摘要】 在工业机器人的控制系统中嵌入机器视觉，可以提高工业机器人作业的智能化，拓宽工业机器人的应用领域。本文以MOTOMAN UP6机器人为基础，利用工业相机、图像采集卡、计算机、变位机等设备，构建了工业机器人视觉伺服控制系统。通过对场景图像的处理与分析，建立了基于图像的视觉伺服控制器，实现了UP6机器人对作业场景下的静态目标和圆周运动目标实时定位与跟踪拾取。本文首先对相机标定的算法进行研究，使用MATLAB工具实现了相机参数的求取，依据标定结果得出图像坐标系与机器人坐标系的映射关系。然后对相机拍摄的场景图像进行分析，通过实验建立起期望图像的特征矩阵。对场景中的目标进行边缘检测、模板匹配等算法处理，得出实时图像特征矩阵。利用图像间的特征信息建立起图像反馈控制模块，设计了静态场景和动态场景两种情况下的视觉伺服控制器。在运动目标跟踪方面，研究了目标作圆周运动的特征，建立了“两心一度”模型，通过实验验证了该算法的正确性。结合卡尔曼滤波算法和全局最近邻预测算法，对视觉伺服控制器的设计进行改进，使得机器人末端执行器实时跟踪定位、拾取作业场景中的目标。通过多次实验和程序调试，完成了一个初级的机器人视觉伺服控制系统的构建。联机实验证明了本文的研究成果。在研究的过程中发现UP6机器人运动控制过程中的不足之处，并提出了解决方法。该系统作为工业自动化的一个环节，在生产应用上有较好的实用意义和参考价值。更多还原

【Abstract】 The industrial robot control system could be embedded in machine vision, which willimprove the intelligence of industrial robots and widen industrial robots applications. The paperconstructs the industrial robot visual servo control system which includes CCD camera, imageacquisition card, computer, positional machine and other equipments. Through the scene ofimage processing and analysis, UP6robot can pick the static targets and trick the circularmotion targets.Firstly, the paper researches the algorithm of camera calibration, uses the MATLAB toachieve the camera parameter. According to the calibration results, which can get the mappingrelationship between graphical coordinate and robot coordinate. Secondly, as the analysis ofimage, the paper establishes a desired image characteristic matrix through the experiments.Using the algorithm of edge detection、template matching, which can get the feature matrix ofthe real-time image and establish the control of image feedback. Then the paper designs visualservo controller of static scene and dynamic scene. In terms of tricking dynamic targets, thepaper studies the circular motion targets and establishes "two hearts and one angular speed"model, through experimental verification to prove the correctness of the algorithm. CombinedKALMAN filtering and the global nearest neighbor algorithm, which can improve theperformance of visual servo controller and control robot to complete the servo works.Through several experiments and program tests, the paper constructs a primary robotvisual servo control system. Online experiments prove the results of this research. In theprocess of research, some deficiencies of robot motion control have been found and thesolutions are proposed. The research of visual servo system for industrial automation is animportant link, which has practical significance and reference value in the production andapplication.更多还原