基于双目视觉的集装箱自动识别定位系统的设计  

Design of auto-locating recognition system for container based on binocular stereo vision

作　　者：梁晓波[1] 程文明[1] 

机构地区：[1]西南交通大学机械工程学院,四川成都610031

出　　处：《机械》2015年第1期7-10,共4页Machinery

基　　金：国家自然科学基金项目(51175442);四川省科技计划项目(11ZC0672);四川省应用基础研究计划项目(2013JY0136)

摘　　要：针对集装箱装卸过程中存在的耗时长、过程复杂和定位不精确等问题,提出了一种基于双目立体视觉识别定位原理的集装箱自动识别系统。系统使用COMS高清摄像机代替人眼采集目标集装箱图像,采用ARM处理器运行算法程序,计算出目标集装箱通三维坐标信息。通过现场实验,结果表明系统可靠适用,所计算测量集装箱位置坐标误差在1%以内,满足自动识别定位要求,具有重要的应用参考价值。In order to solve the problems of the long time, complicated process and inaccurate positioning existed in the process of container handling, a type of auto-locating recognition system for container which based on binocular stereo vision was presented. Instead of human eye, two COMS cameras were used to collect image, finally the 3D coordinates of container were calculated by ARM processor. The actual experimental results show that the auto-locating recognition system for container cranes was reliable and applicable, the relative error of measurement is within 1%, had important application and reference value.

关 键 词：双目立体视觉 自动定位 集装箱 

分 类 号：TP274.5[自动化与计算机技术—检测技术与自动化装置]