基于遗传算法的点云数据配准

在光学非接触三维测量中，复杂对象的重构需要多组测量数据的配准。为此，提出一种基于遗传算法的线扫描点云数据配准方法。曲面线扫描点云数据同一表面的拓扑结构在不同视图下曲率变化趋势相同，根据该性质，利用遗传算法识别两点云数据集的重叠区域，并求解子集的坐标变换矩阵，完成配准。实验结果表明，与ICP算法相比，该方法的运行速度较快，且配准精度较高。

Registration of Point Clouds Data Based on Genetic Algorithm

In the optical non-contact measurement process, the reconstruction of complex object depends on the registering of point clouds. Aiming at this problem, this paper proposes a registration method of point clouds data based on Genetic Algorithm(GA). At different views, according to the same change trend of curvatures of the topology at same surface about point clouds of scanning beam, based on GA, the overlapping region is identified and transition matrix is directly extracted, and the registration is completed. Experimental results show that the method has better registration performance.