• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Journal of Imaging Science and Technology >Fast Registration of Three-Dimensional Laser Scans without Initial Guess
【24h】

Fast Registration of Three-Dimensional Laser Scans without Initial Guess

机译：无需初步猜测即可快速注册三维激光扫描

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

State of the art graph-based 3D simultaneous localization and mapping (SLAM) systems are typically divided into the so-called front-and back end. The front end's task is to align laser scans locally, constructing a graph of the measurements in the process. The back end's task is to optimize this graph in order to find a maximally consistent configuration of the nodes. The well-known iterative closest point (ICP) algorithm is often used to align pairs of laser scans. As it only provides local convergence, it generally requires an approximate guess in order to determine the correct relative transformation. In practice, this guess is often obtained from a robot's odometry. However, in some applications this odometry information is not available and a fast initial guess is needed for real-time operation. Existing feature-based approaches are often unsuitable for real-time operation due to their computational requirements. The authors present a featureless algorithm that is able to compute an approximate transformation between two laser scans quickly, serving as initial guess for ICP. This algorithm is an adaptation to 3D of a 2D correlative scan-matching algorithm by Olson and will be evaluated in the context of 3D SLAM. Comparisons are made to results that use features as they are used in image processing to match images. The authors make some plausible and moderate assumptions on the scan acquisition and experimentally show that these are usually met. The experiments show that the presented algorithm is able to align two 3D laser scans quickly and reliably in the context of mobile robotics and stand-alone laser scanning. (C) 2014 Society for Imaging Science and Technology.
机译：现有技术中基于图形的3D同时定位和映射（SLAM）系统通常分为所谓的前端和后端。前端的任务是在本地对齐激光扫描，以构建过程中的测量图。后端的任务是优化此图，以便找到节点的最大一致配置。众所周知的迭代最近点（ICP）算法通常用于对齐成对的激光扫描。由于它仅提供局部收敛，因此通常需要进行近似猜测才能确定正确的相对变换。实际上，这种猜测通常是从机器人的里程表中获得的。但是，在某些应用中，该测距信息不可用，并且实时操作需要快速的初始猜测。现有的基于特征的方法由于其计算要求，通常不适合实时操作。作者提出了一种无特征的算法，该算法能够快速计算两次激光扫描之间的近似变换，从而可以作为ICP的初始猜测。该算法是Olson对2D相关扫描匹配算法的3D的一种改进，将在3D SLAM的背景下进行评估。对使用在图像处理中匹配图像的功能的结果进行比较。作者对扫描采集做出了一些合理而适度的假设，并通过实验证明了通常可以满足这些假设。实验表明，所提出的算法能够在移动机器人技术和独立激光扫描的情况下快速，可靠地对齐两个3D激光扫描。 （C）2014年影像科学与技术学会。

著录项

  • 来源
    《Journal of Imaging Science and Technology》 |2014年第6期|060403.1-060403.6|共6页
  • 作者

    Neuhaus Frank; Muetzel Andreas; Palulus Dietrich;

  • 作者单位

    Univ Koblenz Landau, Dept Computat Visualist, D-56070 Koblenz, Germany;

    Univ Koblenz Landau, Dept Computat Visualist, D-56070 Koblenz, Germany;

    Univ Koblenz Landau, Dept Computat Visualist, D-56070 Koblenz, Germany;

  • 收录信息
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类
  • 关键词

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号