机器视觉在机器人杂乱工件分拣中的应用

发布时间：2019-04-22 21:19

【摘要】：随着工业自动化的飞速发展，机器人分拣技术被普遍地应用在各类生产线上。传统的机器人分拣大多采用示教方式，对机器人的初始和终止姿态及工件的摆放位置有严格的要求，这样不但分拣速度慢、效率低，而且一旦工件位置发生变化则会导致机器人抓取失败，从而影响生产效率。于是，人们把视线转移到机器视觉领域，将视觉技术与机器人相结合，使机器人拥有类似人眼的识别功能，从而使机器人分拣时更加柔性化、高精化、智能化，对于降低劳动成本、保证产品质量、提高生产效率等方面具有重要意义。如今，机器视觉技术已经被广泛地应用在无损检测、食品包装、医药生产、物流分拣、PCB制图等诸多领域，基于视觉系统引导的机器人技术也将成为未来发展的主要趋势。本文针对从生产线上的杂乱工件堆中分拣出目标工件，提出了一种基于模板匹配的工件视觉识别算法，并以新松公司RH6-A型6轴工业机器人为基础，搭建了基于机器视觉的工业机器人分拣系统，制定了详细的视觉分拣的方案。本文首先进行了工业机器人运动学分析，应用D-H坐标变换法求出了运动学正解及反变换法解出反解，并根据该算法开发了基于OpenGL的机器人仿真系统。其次分析了OpenCV视觉算法库中提供的摄像机模型，并采用张正友标定法进行了摄像机标定，建立了机器人手眼关系，求解出摄像机内、外参数及变换矩阵。然后通过工业相机采集工件图像，，对采集到的图像进行平滑、去噪等预处理，以消除图像中掺杂的噪声，再利用Sobel边缘算子对工件进行边缘检测，得到工件的边缘信息，接着进行图像二值化处理，最后利用Hausdorff距离算法对模板图像和待匹配图像进行相似性度量，完成模板匹配。实验结果表明，基于边缘的模板匹配算法成功地识别匹配出了杂乱工件中的目标工件，然后通过图像的中心矩计算出了工件的中心坐标，并将特征信息反馈给机器人控制器，达到机器人分拣工件的目的。
[Abstract]:With the rapid development of industrial automation, robot sorting technology is widely used in various production lines. The traditional robot sorting mostly adopts the teaching method, which has strict requirements on the initial and terminating posture of the robot and the placement position of the workpiece, so that the sorting speed is slow and the efficiency is low. And once the position of the workpiece changes, it will lead to the failure of the robot grab, thus affecting the production efficiency. Therefore, people turn their sight to the field of machine vision, combine vision technology with robot, so that robot has the recognition function similar to human eye, so that the robot can be more flexible, highly refined and intelligent when sorting. It is of great significance to reduce labor cost, guarantee product quality and improve production efficiency. Nowadays, machine vision technology has been widely used in many fields such as nondestructive testing, food packaging, pharmaceutical production, logistics sorting, PCB drawing and so on. Robot technology based on vision system will also become the main trend of future development. In this paper, a visual recognition algorithm of workpiece based on template matching is proposed, which is based on RH6- A-6 axis industrial robot of Xinsong Company, aiming at sorting out the target workpiece from the cluttered workpiece heap on the production line. An industrial robot sorting system based on machine vision is built, and a detailed visual sorting scheme is formulated. In this paper, the kinematics of industrial robot is analyzed firstly, and the forward and inverse kinematics solutions are obtained by using DH coordinate transformation method. According to this algorithm, the robot simulation system based on OpenGL is developed. Secondly, the camera model provided in the OpenCV vision algorithm library is analyzed, and the camera calibration is carried out by using Zhang Zhengyou calibration method. The hand-eye relationship of the robot is established, and the inner and outer parameters of the camera and the transformation matrix are solved. Then the image of the workpiece is collected by the industrial camera, and the collected image is smoothed and de-noised, so as to eliminate the doping noise in the image. Then, the edge information of the workpiece is obtained by using the Sobel edge operator to detect the edge of the workpiece. Then the binary image processing is carried out. Finally, the Hausdorff distance algorithm is used to measure the similarity between the template image and the image to be matched to complete the template matching. The experimental results show that the edge-based template matching algorithm successfully recognizes and matches the target workpiece in the clutter workpiece, then calculates the center coordinates of the workpiece through the central moment of the image, and feeds the feature information back to the robot controller. To achieve the purpose of robot sorting workpiece.
【学位授予单位】：沈阳工业大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP391.41;TP242

【参考文献】