基于形状内容分析的机器人物料分拣系统  

Robot Material Sorting System Based on Shape Contexts Analysis

作　　者：邹腾跃[1] 唐小琦[1] 宋宝[1] 陈吉红[1] 

机构地区：[1]华中科技大学国家数控系统工程技术研究中心,武汉430074

出　　处：《组合机床与自动化加工技术》2013年第5期48-51,共4页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：国家自然科学基金(50905069);国家科技重大专项(2012ZX04001012);2012国家支撑计划(2012BAF13B01)

摘　　要：文章分析了国内外工业机器人分拣系统的现状后,提出运用形状内容分析方法对物料的分类识别进行强化,以提高分拣的准确性和稳定性。设计了以数控系统、现场总线、伺服驱动单元以及以太网通信链路组成的机器人控制系统硬件结构。控制软件以运行在嵌入式Linux系统上的实时响应程序为基础,涵盖电机驱动、异常检测、运动控制等多个方面,并为上位机图像分析结果的传输提供接口。验证工作基于华中数控HNC-08数控系统和MOTOMAN-SK6机器人展开,通过标定后的摄像机进行物料的定位与分类识别,并通过机器人控制系统进行分拣,实验结果充分证明了系统的准确性和稳定性。The article analyses the current situation of the domestic and foreign industrial robot sorting system. Then, it proposes the shape contexts analysis to strengthen the identification and classification of materials for improving the accuracy and stability of the sorting system. It also designs a hardware struc- ture of the robot control system composed of numerical control systems, fieldbus, servo drive unit and Ethernet communication link. The control software is based on the real-time response program which runs on embedded Linux system, covering motor drive, anomaly detection, motion control and many other aspects with interface for transmission of the image analysis results from host computer. Validation work is based on Huazhong-CNC HNC-08 CNC system and MOTOMAN-SK6 robot. The materials are identi- fied and classified through calibrated digital camera, and then sorted by robot control system. The experi- mental results demonstrate accuracy and stability of the proposed system.

关 键 词：形状内容 目标识别 工业机器人 分拣 机器视觉 

分 类 号：TH165[机械工程—机械制造及自动化] TG65[金属学及工艺—金属切削加工及机床]