摘要
为了解决使用传统示教或离线编程方式的工业机器人无法满足复杂分拣环境作业要求的问题,以史陶比尔工业机器人、Kinect深度相机为硬件基础,搭建了基于机器视觉的工业机器人智能分拣系统平台。通过支持向量机(SVM)实现对分拣物体的学习识别,通过基于采样一致性(SAC-IA)粗配准和迭代最近点(ICP)精配准算法实现对物体位姿点云数据的配准。实验结果表明,搭建的智能分拣系统能够识别分拣物体种类并且获取物体的位姿,拓宽了工业机器人的应用领域。
In order to solve the problem that industrial robots using traditional teaching or offline programming methods can not meet the requirements of complex sorting environment,this paper built industrial robot intelligence based on machine vision. The platform based on St ubli industrial robot and Kinect depth camera was set up.The sorting system platform was used to support vector machine (SVM) to realize the learning and recognition of sorting objects. The registration of the point cloud data of the object is realized by Sampling Consistency (SAC-IA) coarse registration and iterative closest point (ICP) fine registration algorithm. The experimental results show that the intelligent sorting system built in this paper can identify the sorting object type and obtain the object pose and broaden the application field of industrial robots.
作者
徐青青
XU Qing-qing(College of Mechanical and Electrical Engineering,Suqian College,Suqian 223800,China)
出处
《仪表技术与传感器》
CSCD
北大核心
2019年第8期92-95,100,共5页
Instrument Technique and Sensor
基金
江苏省高等学校自然科学研究项目资助(17KJB470013)
江苏省高校品牌专业建设工程资助项目(PPZY2015C252)
宿迁市科技计划项目(Z2018097,S201711)
宿迁学院科研基金项目(2016KY08)
关键词
工业机器人
机器学习
物体识别
点云配准
物体分拣
industrial robot
machine learning
object recognition
point cloud registration
object sorting
