四自由度点云初始配准算法对比

Comparison of Initial Registration Algorithms for Four-Degree-of-Freedom Point Cloud

针对更贴合实际点云数据处理的四自由度点云初始算法,进行了背景的介绍,选取了目前主流的5种四自由度点云初始配准算法,分别为RANSAC算法、LM算法、Bn B算法、FMP-Bn B算法和K-4PCS算法;对2组室内场景点云数据和2组室外场景点云数据进行配准,对配准结果从精度和效率两方面进行了对比分析。实验结果表明,不同的算法适用的数据类型不同,在进行实际的数据初始配准时,可依据数据类型选取合适的算法。

Aiming at the four-degree-of-freedom point cloud initial algorithm that is more suitable for actual point cloud data processing,this paper introduces the background and selects the current mainstream five four-degree-of-freedom point cloud initial registration algorithms,namely RANSAC algorithm,LM algorithm,The BnB algorithm,FMP-BnB algorithm and K-4PCS algorithm are used to register two sets of indoor scene point cloud data and two sets of outdoor scene point cloud data,and the registration results are compared and analyzed in terms of accuracy and efficiency.The experimental results show that different algorithms are applicable to different data types.In the actual initial data registration,the appropriate algorithm can be selected according to the data type.