基于视觉定位的六轴工业机器人搬运系统设计与研究

发布时间：2018-03-13 08:39

本文选题：搬运机器人　切入点：相机标定　出处：《合肥工业大学》2017年硕士论文　论文类型：学位论文

【摘要】：搬运机器人在工业生产中应用广泛,然而传统的示教方法无法应对需要搬运的物体位姿不定的情况,因此引入视觉技术对目标物体进行识别和定位,引导搬运机器人完成搬运任务。以空调压缩机搬运上线为目标任务,研究先后涉及机器人运动学、相机标定与手眼标定、图像处理与目标识别等方面,最后根据压缩机自动搬运上线的任务要求完成基于视觉定位的机器人搬运系统的设计应用。在机器人运动学研究中,以RS50N型六轴工业机器人为研究对象,分析了该型机器人的结构及基本参数,建立了连杆坐标系,列出相应D-H参数,之后根据D-H参数进行了机器人运动学正逆解的求解,为机器人手眼标定和搬运系统正确定位目标提供技术支持。在相机标定研究中,分析了相机基本数学模型,基于张正友法和MATLAB摄像机标定工具箱进行了相机标定实验,求取了相机内参数和相应的外参数;在手眼标定研究中,对Eye-in-Hand视觉系统进行了手眼标定研究,推导了手眼标定基本关系式AX=XB,选取基于矩阵直积的方法进行标定矩阵的求解,为减少噪声影响,进一步研究通过整体最小二乘法实现手眼标定关系矩阵的精确求解,最后通过实验完成了Eye-in-Hand视觉系统的手眼标定。在图像处理与目标识别研究中,以MATLAB图像处理工具箱为平台,以拍摄的一幅图像为处理对象,分析研究了图像预处理、边缘检测方法;并针对图像中的圆形目标物体的识别,提出基于圆Hough变换和最小二乘法相结合的圆形目标检测定位方法,实验表明该方法识别效率和定位精度均较高。最后根据以上研究成果和压缩机自动搬运上线的任务要求,构建了基于机器视觉和RS50N六轴工业机器人的压缩机自动搬运系统。其中,设计并优化了机器人末端夹爪机构。经实验验证该系统能够出色完成工作任务。
[Abstract]:The robot is widely used in industrial production, but the traditional display can not cope with the need to pose object handling uncertain situation teaching method, so the introduction of visual technology to identify and locate the target object, guided robot handling tasks. By air conditioning compressor on-line handling task as the goal, research has involved the robot kinematics, camera calibration with the hand eye calibration, image processing and target recognition. Finally, based on the compressor automatic handling on-line tasks required to complete the design and application of robot handling system based on visual positioning. In the kinematics of the robot, with RS50N type six axis industrial robot as the research object, analyzes the structure of the robot and basic parameters, establish linkage coordinate system, lists the corresponding D-H parameters, according to the D-H parameters of the robot kinematics inverse solution, for the robot Hand eye calibration and provide technical support for handling system the correct positioning of target. In the research of camera calibration, analysis of the basic mathematical model of the camera, the camera calibration experiment of Zhang's method and MATLAB camera calibration toolbox based on calculated camera parameters and the corresponding parameters; in the study of hand eye calibration, vision system of Eye-in-Hand calibration of hand eye derived hand eye calibration basic formula AX=XB, were selected for the calibration matrix method based on matrix direct product, to reduce noise effects, further research through the overall least squares implementation of hand eye calibration accurate solution matrix, finally completed the Eye-in-Hand calibration of vision system in image processing and hand eye. Study on recognition, using MATLAB image processing toolbox for the platform, in an image taken as treatment object, research the image Preprocessing, edge detection and recognition methods; for a circular object in the image of the proposed circular target detection and localization method of circular Hough transform and least square method based on the combination of experiments show that the method recognition efficiency and positioning accuracy are high. According to the above research results and the compressor on-line automatic handling tasks, automatic construction handling compressor machine vision and RS50N six axis industrial robot system based on the design and optimization of the robot gripper mechanism. The experimental results show that the proposed system can complete work tasks.

【学位授予单位】：合肥工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TP242

【参考文献】