期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：2008
卷号：XXXVII Part B3b
页码：179-184
出版社：Copernicus Publications
摘要：Generation of digital terrain model on the basis of airborne laser scanning data requires extracting exclusively these points from the raw point cloud which are reflections of a laser beam from the ground. This task is performed mainly automatically using specialist software for classification or filtration of laser scanning points. Various filtration methods have been suggested but neither of the proposed algorithms guarantees 100% efficiency. In this paper, the algorithm based on surface energy minimization is discussed. Total energy of the surface is the sum of internal and external energy. Internal energy describes geometrical properties of the modelled surface and for the introduced flakes model it equals the weighted sum of surface membrane kernel and surface thin plate kernel. External energy describes difference between estimated active surface and survey data and depends on the difference between the measured height and approximated height. As a result of total surface energy minimization, the active surface is adjusted to the terrain surface. The variational problem that occurs in the surface energy minimization was solved using the direct method (Ritz method). Flakes method was tested on authentic laser scanning data, which additionally contained reference data, i.e. correctly classified terrain points and object points. Comparison of reference data and sets of points obtained in filtration facilitated determination of the percentage values of filtering errors. The results confirmed that flakes method is effective. Filtering correctness value is similar to results which were obtained using other methods
