基于 3D.Harris关键点检测结合改进 ICP的姿态估计算法
CSTR:
作者:
作者单位:

(东南大学仪器科学与工程学院,南京 210096)

作者简介:

陈宇超(1999-),男,浙江省宁波市人,硕士研究生,主要研究方向为姿态估计算法;蔡体菁(1961-),男,浙江省宁波市人,博士,教授,博士生导师,主要从事惯性导航、组合导航和重力测量技术的研究。

通讯作者:

中图分类号:

TP391. 4;TN958. 98

基金项目:


Attitude Estimation Algorithm Using 3D-Harris Keypoint Detection Combined With Improved ICP
Author:
Affiliation:

(School of Instrument Science and Engineering,Southeast University,Nanjing 210096,CHN)

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    针对传统激光雷达姿态估计过程中原始点云存在离群噪声点、配准易陷入局部最优、计算效率低的问题,文章提出一种基于 3D-Harris关键点与改进迭代最近点(ICP)相结合的三维点云配准算法。该算法首先采用体素滤波与统计滤波进行点云预处理,以去除异常噪声点;随后,利用 3D-Harris关键点检测算法,缩减姿态估计过程中的对应点搜索空间,提升配准效率;在粗配准阶段,基于快速点特征直方图特征描述的采样一致性算法,对关键点进行特征描述并提供初始旋转平移矩阵,以优化精配准的初始值,提升姿态估计精度;在精配准过程中,通过 Kd-tree的近邻搜索,优化 ICP算法执行效率。实验结果表明,所提出的算法在配准精度和计算效率方面均优于传统方法,具有较高的应用价值。

    Abstract:

    To address challenges such as outlier noise,sensitivity to local optima during registration,and the low efficiency of traditional LiDAR-based attitude estimation using raw point clouds,this study proposes a novel 3D point cloud registration algorithm. The method integrates 3D-Harris keypoint detection with an improved Iterative Closest Point(ICP)algorithm. Pre-processing combines voxel grid filtering and statistical outlier removal to eliminate abnormal noise points. Subsequently,3D-Harris keypoint detection is applied to reduce the search space for correspondence, enhancing computational efficiency. For coarse registration,a sample consensus initial alignment algorithm based on Fast Point Feature Histograms is employed to describe keypoint features and estimate an initial rotation and translation matrix. This initialization improves the accuracy and convergence of the fine registration stage. The fine registration process utilizes a Kd-tree-based nearest neighbor search to accelerate the ICP algorithm. Experimental results show that the proposed method achieves higher registration accuracy and computational efficiency than traditional approaches,demonstrating strong practical applicability.

    参考文献
    相似文献
    引证文献
引用本文

陈宇超,蔡体菁.基于 3D. Harris关键点检测结合改进 ICP的姿态估计算法[J].半导体光电,2025,46(5):919-927. CHEN Yuchao, CAI Tijing. Attitude Estimation Algorithm Using 3D-Harris Keypoint Detection Combined With Improved ICP[J].,2025,46(5):919-927.

复制
分享
相关视频

文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2025-04-30
  • 最后修改日期:
  • 录用日期:
  • 在线发布日期: 2025-10-31
  • 出版日期:
文章二维码