中国安全科学学报 ›› 2019, Vol. 29 ›› Issue (7): 117-122.doi: 10.16265/j.cnki.issn1003-3033.2019.07.019

• 安全工程技术科学 • 上一篇    下一篇

轨道式攀爬机器人系统设计与图像处理

秦海伟1, 陆从飞2 高级工程师, 蒋永生3 研究员, 周旭3 高级工程师, 程斌**1 副研究员   

  1. 1 上海交通大学 船舶海洋与建筑工程学院,上海 200240;
    2 中交公路规划设计院有限公司,北京 100088;
    3 贵州高速公路集团有限公司,贵州 贵阳 550009
  • 收稿日期:2019-04-10 修回日期:2019-06-12 出版日期:2019-07-28 发布日期:2020-10-21
  • 通讯作者: ** 程 斌(1979—),男,江西横峰人,博士,副研究员,主要从事钢桥疲劳与断裂、FRP复合材料桥梁、智慧桥梁结构方面的研究。E-mail:cheng_bin@sjtu.edu.cn。
  • 作者简介:秦海伟 (1995—),男,江苏徐州人,硕士研究生,主要研究方向为桥梁检测与养护技术。 E-mail:qinhaiwei1995@126.com。
  • 基金资助:
    贵州省交通运输厅科技项目(2017-121-044,2019-122-004)。

Design of orbital climbing robot system and image processing

QIN Haiwei1, LU Congfei2, JIANG Yongsheng3, ZHOU Xu3, CHENG Bin1   

  1. 1 School of Naval Architecture, Ocean & Civil Engineering, Shanghai Jiaotong University, Shanghai 200240, China;
    2 Research Institute of Highway Ministry of Transport,Beijing 100088, China;
    3 Guizhou Expressway Group Company Limited, Guiyang Guizhou 550009, China
  • Received:2019-04-10 Revised:2019-06-12 Online:2019-07-28 Published:2020-10-21

摘要: 为解决桥塔、桥墩等桥梁高耸结构物检测耗时长、成本高、精度差的问题,研发一套轨道式攀爬机器人系统,并提出一种检测图像处理方法。该系统由机械平台、视觉平台、控制平台等3大部分组成,将视觉平台搭载在机械平台轨道上,并使用控制平台远程控制机械平台及视觉平台工作,实现结构物表面图像的间隔式、连续化采集;根据图像的位置信息,结合尺度不变特征转换(SIFT)算法拼接检测图像,生成桥梁全景图像。通过搭建带有裂缝图像背景的试验平台,进行样机试验,检验该系统的性能。结果表明:相比于现有的检测方法,轨道式攀爬机器人系统的检测效果更加快速稳定,并能通过图像处理方法生成全景图像,其实施将有助于提高桥梁检测与维护的水平。

关键词: 攀爬机器人, 平台协同, 构件设计, 图像拼接, 尺度不变特征转换(SIFT)

Abstract: In order to solve the problem that current bridge inspection for high-rise structures such as bridge pylons and piers is time-consuming, costly and inaccurate, a track-type climbing robot system was developed and a processing method of detection images was proposed. The system consists of three parts: mechanical platform, visual platform and control platform. The visual platform is placed on the mechanical platform and both of them are remotely controlled by the control platform to capture the interval and continuous surface images of the structure. The SIFT algorithm is used to splice the detected images according to the position information to generate the bridge panorama. The prototype experiment was conducted by setting up an experimental platform with a cracked image background. The results show that compared with the existing detection methods, the designed robot system is more rapid and stable, and the panorama can be generated by the image processing method, which will help improve the level of bridge detection and maintenance.

Key words: orbital robot, platform collaboration, component design, image stitching, scale-invariant feature transform(SIFT)

中图分类号: