首页 | 本学科首页   官方微博 | 高级检索  
     检索      

西红柿采摘机器人视觉系统的研究
引用本文:郭凯敏,崔天时,张桢,郭志强,朱铁欣,谢学刚.西红柿采摘机器人视觉系统的研究[J].农机化研究,2016(12):217-221,226.
作者姓名:郭凯敏  崔天时  张桢  郭志强  朱铁欣  谢学刚
作者单位:东北农业大学,哈尔滨,150030
基金项目:国家“863计划”项目(810028)
摘    要:针对果实采摘机器人果实识别率低的问题,设计了一组用于西红柿识别和定位的双目立体视觉系统,为机器人的采摘作业提供更有利的条件。为此,采用Bumblebee双目立体视觉系统,基于成熟果实与植株颜色特征的差异进行图像分割,来识别成熟的西红柿;在完成相机标定、特征点提取和特征点匹配的基础上,通过三维空间定位获取果实的三维坐标。实验结果表明:该系统果实识别的整个过程平均耗时150ms,对成熟西红柿的识别率达到99%,测试误差在10mm以内,能够较好地满足西红柿采摘工作的要求。

关 键 词:农业机器人  双目立体视觉  相机标定  特征点提取  特征点匹配  三维重建  定位

Research on Tomato-harvesting Robot Visual System
Guo Kaimin;Cui Tianshi;Zhang Zhen;Guo Zhiqiang;Zhu Tiexin;Xie Xuegang.Research on Tomato-harvesting Robot Visual System[J].Journal of Agricultural Mechanization Research,2016(12):217-221,226.
Authors:Guo Kaimin;Cui Tianshi;Zhang Zhen;Guo Zhiqiang;Zhu Tiexin;Xie Xuegang
Institution:Guo Kaimin;Cui Tianshi;Zhang Zhen;Guo Zhiqiang;Zhu Tiexin;Xie Xuegang;Northeast Agricultural University;
Abstract:Aimed at the problem of low recognition rate of fruit picking robots ,design a set of binocular stereo vision sys-tem is used to identify and locate tomato , provide more favorable conditions of picking robot operation .The Bumblebee binocular stereo vision system , the identification of visual system is developed through the difference between the color characteristics of ripe tomatoes and plant .after camera calibration , feature point extraction and feature points matching , the space coordinates is obtained through 3 D reconstruction .The experiment results show that the identification rate of the vision system can reach 99%,and identification average time is about 0.15s,the absolute error of distance-measuring can be controlled within 10mm.It can meet the needs of practical work very well .
Keywords:agricultural robot  binocular stereo vision  camera calibration  feature point extraction  feature points matc-hing  3 dreconstruction  location
本文献已被 CNKI 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号