自定义聚类中心点的快速K-means聚类点云精简算法  

Fast K-means Clustering Point Cloud Reduction Algorithm for Custom Clustering Center Points

作　　者：王世刚 关红利 

机构地区：[1]广西科技大学自动化学院,广西柳州545616

出　　处：《工业控制计算机》2024年第8期123-125,共3页Industrial Control Computer

基　　金：广西科技基地和人才专项(桂科AD22080004)。

摘　　要：针对传统K-means算法在随机选取聚类中心点出现聚类失败及点云数据重建时在相对平坦的区域出现孔洞的问题,提出一种GK-means的改进聚类算法对点云数据进行精简。该算法首先规定数值K作为最终聚类个数的限定,然后对选点策略进行改进,采用最远点采样选取聚类中心,对簇进行细分,计算所有点到聚类中心的欧氏距离,获取最小值所在的位置,放进最小距离所在的簇。实验结果表明:改进后的K-means算法能够使算法成功的概率提高且运行速度较快,对点云进行精简时,特征区域完整地保留了点云模型的细节特征,重建结果具有较高的光顺性。An improved clustering algorithm of GK-means is proposed to simplify point cloud data in this paper.The algorithm first stipulates the value K as the limit of the final number of clusters,and then improves the point selection strategy,selects the cluster center with the farthest point sampling,subdivides the cluster,calculates the Euclidean distance from the point to the cluster center,obtains the position of the minimum value,and puts it into the cluster where the minimum distance is located.The experimental results show that the improved K-means algorithm can improve the success probability of the algorithm and run faster.When the point cloud is simplified,the feature region can retain the detailed features of the point cloud model completely,and the reconstructed results have high smoothness.

关 键 词：聚类中心 迭代 GK-means算法 点云精简 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]