基于鲸鱼优化算法的点云配准算法

Point Cloud Registration Algorithm Based on Whale Optimization Algorithm

摘要: 点云配准的核心是通过求解源点云与目标点云的最优刚性变换矩阵实现两者对齐，其性能直接影响同步定位与测绘、三维重建等领域的应用效果。然而，传统算法存在配准精度低、耗时久及鲁棒性不足的问题。为此，本文提出一种哈希函数与鲸鱼优化算法(WOA)协同的点云配准方法：利用哈希函数对高维点云特征进行低维映射，快速筛选潜在匹配点对；再通过鲸鱼优化算法对粗配准矩阵进行迭代优化，提升变换参数精度。以斯坦福公共点云数据集的Bunny、Dragon、Buddha和Armadillo模型为实验对象，将所提算法与SAC-IA、FPCS、PCA三种传统算法对比。实验结果表明，本文算法在配准时间与精度综合性能上均显著优于对比算法。

Abstract: The core of point cloud registration is to align the source point cloud and target point cloud by solving the optimal rigid transformation matrix, and its performance directly affects the application effects in fields such as simultaneous localization and mapping (SLAM) and 3D reconstruction. However, traditional algorithms suffer from problems of low registration accuracy, long time consumption, and insufficient robustness. To address these issues, this paper proposes a point cloud registration method that collaborates hash function with Whale Optimization Algorithm (WOA): the hash function is used to perform low-dimensional mapping of high-dimensional point cloud features, enabling rapid screening of potential matching point pairs; then the Whale Optimization Algorithm is adopted to iteratively optimize the coarse registration matrix, improving the accuracy of transformation parameters. Experiments are conducted on the Bunny, Dragon, Buddha, and Armadillo models from the Stanford public point cloud dataset, and the proposed algorithm is compared with three traditional algorithms (SAC-IA, FPCS, and PCA). The experimental results show that the proposed algorithm significantly outperforms the comparison algorithms in terms of the comprehensive performance of registration time and accuracy.