[1]邢 静,闫克丁,高俊钗.基于 IMU 点云特征跟踪一致性的 ICP 配准技术[J].机械与电子,2022,(03):3-7.
　XING Jing,YAN Keding,GAO Junchai.ICP Registration Technology Based on IMU Point Cloud Feature Tracking Consistency[J].Machinery & Electronics,2022,(03):3-7.
点击复制

基于 IMU 点云特征跟踪一致性的 ICP 配准技术()
分享到：

机械与电子[ISSN:1001-2257/CN:52-1052/TH]

卷:
期数:
2022年03期
页码:
3-7
栏目:
设计与研究
出版日期:
2022-03-25

文章信息/Info

Title:
ICP Registration Technology Based on IMU Point Cloud Feature Tracking Consistency
文章编号:
1001-2257 ( 2022 ) 03-0003-05
作者:
邢 静 1 ; 闫克丁 2 ; 高俊钗 2
1. 西安培华学院智能科学与信息工程学院,陕西 西安 710125 ;2. 西安工业大学电子信息工程学院,陕西 西安 710021
Author(s):
XING Jing1 ;  YAN Keding2 ;  GAO Junchai2
(1.School of Intelligent Science and Information Engineering , Xi ’ an Peihua University , Xi ’ an 710125 , China ; 2.School of Electronic and Information Engineering , Xi ’ an Technological University , Xi ’ an 710021 , China )
关键词:
IMU ;  NARF 关键点;  ICP 配准;  FPFH 特征跟踪
Keywords:
IMU ;  NARF key point ;  ICP registration ;  FPFH feature tracking
分类号:
TP391.41
文献标志码:
A
摘要:
针对点云非重叠区域较大时,由于场景中存在相似区域和平滑区域, ICP 配准易陷入局部极小值的问题,设计了一种初始配准与精配准结合的快速准确配准算法。首先,结合点云深度图像进行网格分区,基于蒙特卡罗方法随机选取 20 个 NARF 关键点,利用 IMU 惯导跟踪关键点的 FPFH 特征,并根据跟踪矢量的一致性剔除误匹配,快速计算初始配准矩阵;然后,利用所有关键点进行最近邻跟踪匹配,并根据跟踪矢量的一致性剔除误匹配,提高了 ICP 配准的估计精度。通过 Bunny 兔和 NYUv2 数据集将该算法与 ICP 算法进行对比,验证了该算法能够有效地提高点云配准效率和精度。
Abstract:
Due to the existence of similar and smooth regions in the large non-overlapping region of point cloud , ICP registration is prone to fall into local minimum.Combining initial registration with precision registration , a fast and accurate registration algorithm is designed.Firstly , the FPFH feature is tracked by Inertial Navigation System ( IMU ) based on 20 NARF key points selected randomly by Monte Carlo method in the grid partitioning of depth images , and the false matches are eliminated according to the consistency of tracking vectors , the initial registration matrix is computed quickly.And then the nearest neighbor tracking match is used to all the key points , the false matches are eliminated according to the consistency of tracking vectors , which improve the estimation accuracy of ICP registration.The proposed algorithm is compared with the ICP algorithm through the data set of Bunny rabbit and NYUv2.The results show that the proposed algorithm improves the efficiency and accuracy of point cloud registration effectively.

参考文献/References:

[ 1 ] RUSU R B , BLODOW N , BEETZ M.Fast point feature histograms( FPFH ) for 3D registration[ C ]// 2009 IEEE International Conference on Robotics and Automation , 2009 : 3212-3217.

[ 2 ] SERG B , JITENDRA M , JAN P , et al.Shape matching and object recognition using shape contexts [ J ] .Transactions onpattern analysis and machine intelligence ,2002 , 24 ( 4 ): 509-522.
[ 3 ] BESL P J , MCKAYN D.A method for registration of 3D shape [ J ] .IEEE Transactions on pattern analysis and machine intelligence , 1992 , 14 ( 2 ): 239-256.
[ 4 ] BIBER P , STRASSER W.The normal distributions transform : a new approach to laser scan matching[ C ]// 2003 IEEE / RSJ International Conference on Intelligent Robots and Systems ( IROS 2003 ), 2003 : 2743-2748.
[ 5 ] 宗晓萍,吴岩 . 基于 FPFH 特征的 ICP 点云配准的改进算法[ J ] . 工业控制计算机, 2021 , 34 ( 7 ): 106-108.
[ 6 ] RUSU R B , BLODOW N , BEETZ M.Fast Point Feature Histograms ( FPFH ) for 3D registration [ C ]// 2009 IEEE International Conference on Robotics and Automation , 2009 : 3212-3217.
[ 7 ] 石峰源,张春明,姜丽辉,等 . 采用主成分分析的 ICP 算法优化与验证[ J / OL ] . 激光与光电子学进展, 2021 , 58( 20 ): 1-17 [ 2021-10-01 ] .https :// kns.cnki.net / kcms / detail / 31.1690.tn.20210712.1708.066.html.
[ 8 ] 荆路,武斌,方锡禄 . 基于 SIFT 特征点结合 ICP 的点云配准方法[ J ] . 激光与红外, 2021 , 51 ( 7 ): 944-950.
[ 9 ] LOWE D G.Distinctive image features from scale invariant keypoints [ J ] .International journal of computer vision , 2004 , 60 ( 2 ): 91-110.
[ 10 ] STEDER B , RUSU R B , KONOLIGE K , et al. NARF : 3D range image features for object recognition [ C ]// Workshop on Defining and Solving Realistic Perception Problems in Personal Robotics at the IEEE / RSJ Int.Conf.on Intelligent Robots and Systems( IROS ), 2010 , 44 : 1-2.

备注/Memo

备注/Memo:
收稿日期: 2021-10-08
基金项目:国家自然科学基金资助项目( 11804263 );西安培华学院校级科研项目( PHKT2006 )
作者简介:邢 静 ( 1987- ),女,江苏丰县人,硕士,讲师,研究方向为光电检测和图像处理;闫克丁 ( 1983- ),男,河北保定人,博士,副教授,研究方向为目标光学特性研究、检测技术与自动化等;高俊钗 ( 1971- ),女,河北石家庄人,博士,副教授,研究方向为计算机视觉与模式识别。
更新日期/Last Update: 2022-03-23