计算机与现代化

• 图像处理 • 上一篇    下一篇

基于地面特征块匹配的机器视觉#br# 用于平整地面AGV行驶轨迹计算方法

  

  1. 1.清华大学深圳研究生院,广东深圳518057;2.哈尔滨工业大学深圳研究生院,广东深圳518055
  • 收稿日期:2016-02-23 出版日期:2016-10-15 发布日期:2016-10-14
  • 作者简介:王凡(1985-),男,辽宁大连人,清华大学深圳研究生院高级工程师,本科,研究方向:计算机视觉与机器学习; 邢彦文(1984-),男,河北承德人,哈尔滨工业大学深圳研究生院高级工程师,硕士 ,研究方向:计算机视觉与机器学习。

Machine Vision AGV Indoor Navigation Method Based on Ground Feature Block Matching

  1. 1. Graduate School at Shenzhen, Tsinghua University, Shenzhen 518057, China;

    2. Harbin Institute of Technology Shenzhen Graduate, Shenzhen 518055, China
  • Received:2016-02-23 Online:2016-10-15 Published:2016-10-14

摘要:


摘要:利用机器视觉技术检测地面特征,将块匹配算法用于AGV导航。首先,该方法对每2帧图像的重叠部分进行识别;然后,利用块匹配算法寻找2幅不同帧图像的重叠部分;最后,通过计算重叠部分图
像在前一帧的偏移量来计算AGV移动的距离,从而完成AGV行驶轨迹的计算。实验结果表明,样车系统在不同的地面情况下达到了预期的效果,验证了该方法的有效性。

关键词: 自动导引运输车, 机器视觉, 图像处理, 块匹配, 室内导航

Abstract:

A machine vision technology is used to detect the ground feature, based on block matching algorithm for AGV navigation. First, the overlap of each two frame images is
identified. Second, the overlap of two images of different frames is found by using feature block matching algorithm. Last, through calculating the overlap image in the previous
frame offset, the distance of the car moving is calculated. So far, the calculation of the cars track is completed. Experiment results show that prototype system of this
method reached the anticipated result under different conditions of ground, and the method is proved to be effective.

Key words: AGV (Automated Guided Vehicle), machine vision, image processing, block matching, indoor navigation

中图分类号: