Algorithm for Characteristic Line Extraction of Point Cloud Slicing Data Based on Linear Geometry Primitives
D-J. Cai, D-J. Yue, Z-K. Zhen, W-L. Xing and S. Yue
Three-dimensional (3-D) laser scanning technology provides an important technical measurement for the surfaces of objects; however, how to quickly and accurately extract the characteristic lines of entities from the mass point cloud is still an urgent problem to solve. Aiming at the problems of previous algorithms, such as redundancy calculation, large limitation, and rough characteristic-line extraction, this paper proposes a fast characteristic-line extraction algorithm of point cloud slicing data based on linear geometry primitives. Experimental results show that the proposed algorithm can generate the characteristic lines rapidly while simultaneously ensuring its accuracy and smoothness, and it has been well applied in the two-dimensional contour line extraction of rural buildings.
Keywords: Terrestrial laser scanner, laser point cloud, mass data, parameter space, statistical analysis, characteristic lines