工业机器人视觉伺服控制方法研究

发布时间：2018-02-09 21:18

本文关键词： 视觉伺服工业机器人图像处理图像雅可比矩阵　出处：《长春工业大学》2017年硕士论文　论文类型：学位论文

【摘要】：机器人视觉伺服是机器视觉领域的热门课题,涉及机器人学、控制理论、图像处理等多门学科知识。视觉机器人具有自主分析环境信息和执行任务的能力,广泛应用于航天、国防和工业自动化领域。本文以六自由度工业机器人为研究对象,研究基于图像的视觉伺服控制系统,通过图像采集获取图像信息后完成图像处理,利用图像特征作为反馈信号,并对图像雅可比矩阵进行辨识,利用辨识的结果应用在视觉伺服控制器中,完成了MATLAB仿真和实验台调试工作。主要研究内容如下:首先,详细分析了机器人视觉伺服控制系统的组成,阐述摄像机模型和机器人运动学分析,并对视觉伺服系统的类别进行介绍,比较基于位置和图像的视觉伺服系统。其次,对目标图像进行图像处理和分析,包括图像采集、图像预处理、图像边缘检测和目标特征提取。在预处理过程中高斯滤波方法对消除环境噪声有着较好的效果。通过几种图像边缘检测方法对比分析,采用Canny边缘检测算子对图像进行边缘检测,利用图像矩的方法来对目标特征进行提取。再次,针对基于雅可比矩阵逆的视觉伺服控制方法进行了研究。为了提高系统响应的实时性,能对速度较快的目标进行有效的跟踪,提出一种带预测补偿的逆雅可比矩阵PI控制算法,并且采用扩展H?粒子滤波算法对图像雅可比矩阵进行在线辨识,与卡尔曼滤波算法进行对比实验研究,仿真实验结果验证视觉伺服控制方法可行性。最后,构建视觉伺服控制系统实验平台,完成了目标识别和目标跟踪实验。采用C++来设计视觉伺服控制系统界面,通过对运动目标的轨迹跟踪实验验证了本文的控制方法在实际工程应用中的有效性,实验结果表明基于图像的视觉伺服系统能够达到预定效果。
[Abstract]:Robot vision servo is a hot topic in the field of machine vision, involving robotics, control theory, image processing and other subjects. Visual robot has the ability to analyze environmental information and perform tasks independently, and it is widely used in spaceflight. In the field of national defense and industrial automation. In this paper, a visual servo control system based on image is studied, which takes six degrees of freedom industrial robot as the research object. The image feature is used as feedback signal, and the image Jacobian matrix is identified. The result of identification is applied to the visual servo controller. The MATLAB simulation and experiment bench debugging are completed. The main research contents are as follows: first, The composition of robot visual servo control system is analyzed in detail, the camera model and robot kinematics analysis are expounded, the categories of visual servo system are introduced, and the visual servo system based on position and image is compared. Secondly, Image processing and analysis of the target image, including image acquisition, image preprocessing, Image edge detection and target feature extraction. Gao Si filtering method has a good effect on eliminating environmental noise in the process of preprocessing. The comparison and analysis of several image edge detection methods, The Canny edge detection operator is used to detect the edge of the image, and the method of image moment is used to extract the target feature. The visual servo control method based on Jacobian matrix inverse is studied. In order to improve the real-time response of the system and to track the faster target effectively, an inverse Jacobian matrix Pi control algorithm with predictive compensation is proposed. And use extended H? The particle filter algorithm is used to identify the image Jacobian matrix online, and compared with the Kalman filter algorithm. The simulation results verify the feasibility of the visual servo control method. Finally, a visual servo control system experimental platform is constructed. The interface of visual servo control system is designed by using C, and the effectiveness of the control method in practical engineering is verified by the trajectory tracking experiment of moving target. Experimental results show that the image based visual servo system can achieve a predetermined effect.
【学位授予单位】：长春工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TP242

【相似文献】