ICP算法在多波束点云条带自动配准中的应用

Application of ICP algorithm in automatic registration of multi-beam point cloud strips

针对当前多波束自动匹配过程仍需借助人工干预,并且处理过程中特征点难以确定的问题,提出了一种基于航带法的多波束连接方式,并且引入迭代最邻近算法(iterative closest point,ICP)进行点云条带匹配。利用ICP算法刚性不变原理,建立源点云与目标点云之间的对应关系,利用公共区域点云,采用奇异值分解,求解相邻条带之间的旋转参数和平移参数。将配准后的相邻条带利用航带法进行连接,完成测区内多波束点云的自动匹配和连接。对实验测区数据进行定性定量分析,采用距离阈值和迭代阈值验证本文算法。通过对比不同地形匹配效果,验证ICP自动匹配算法具有一定的稳健性和适用性。

Manual intervention is still needed and the feature points are difficult to determine in the process for the current multi-beam automatic matching process.This paper proposes a multi-beam connection method based on the track method and introduces the Iterative Closest Point(ICP)algorithm for point cloud strip matching.The corresponding relationship between the source point cloud and the target point cloud is established by the rigid invariance principle of the ICP algorithm.Singular value decomposition is used to solve the rotation parameters and translation parameters between adjacent strips by the common area point cloud.The registered adjacent strips are connected by the track to complete the automatic matching and connection of multi-beam point clouds in the survey area.Qualitative and quantitative analysis of the experimental survey area data is carried out and the distance threshold and iterative threshold are used to verify the algorithm in this paper to compare the effects of different terrain matching and to verify that the ICP automatic matching algorithm has certain robustness and applicability.