分带K-均值聚类的平面标靶定位

Planar Target Location Based on the Zoning K-means Clustering

提出了一种分带K-均值聚类的平面标靶定位方法。根据标靶与测站距离的限制条件,推导了较大噪声点的剔除公式,在整体最小二乘拟合平面的基础上增加了噪声点二次剔除的方法,对经过噪声点剔除的点云数据进行分带、聚类处理。同时,对每一带的聚类中心进行均值化处理,得到每一带的中心点,通过求取不同带中心的均值来确定标靶中心点。实验结果表明,分带K-均值聚类的平面标靶定位模型较适合于平面标靶同名点确定。

This paper presents planar target location methods based on zoning K-means clustering aiming at the impact of the noise points and missing data to planar target. The formula of elimination to larger noise points is derived based on the distance constraints between the target and the station. The second elimination of noise points is added to the paper on the basis of the Total Least-squares fit planar. The zoning and cluster of point cloud data is conducted after removing noise points, and the center of each zoning is determined by the mean processing of cluster center of each zoning. The target center is determined by calculating the mean of different zoning center. The experimental results demonstrate that the planar target location model based on the zoning K-means clustering is suited for the determined of the planar target conjugate.