期刊名称：Archiwum Fotogrametrii, Kartografii i Teledetekcji
印刷版ISSN：2083-2214
出版年度：2007
卷号：17
语种：English
出版社：Main Board of Association of Polish Surveyors
摘要：In the process of creating digital terrain model from airborne laser scanning data,there is a need (a necessity) to extract,from the raw points cloud,only those points which are the reflections of laser beam from the ground. This task is performed mainly automatically,using specialized software for data classification or filtering. In the present paper,and algorithm based on surface energy minimisation was presented. The total energy of surface,is the sum of an internal and external energy. Internal energy describes geometrical properties of modelled surface and,in the flakes model,it is a weighted sum of surface membrane kernel and surface thin plate kernel. External energy describes difference between estimated active surface and measured data and depends on the measured height and approximated height. As a result of total surface energy minimisation,active surface is “matched” with the terrain surface. The variation problem,which occurs in the task of surface energy minimisation,was solved using direct method (Ritz method). Numeric tests were carried out on the real scanning data that contained referenced data in the form of correctly classified ground and object points. Throughout referenced data,the evaluation of presented filtering method correctness could be estimated. As a result of comparison of the referenced data with the sets of points,after filtering the percentage values of filtering,errors were calculated. The results achieved confirmed that flakes method is effective – the filtering correctness value is similar to the values obtained using other methods,and amounts to above 90%.
其他摘要：W procesie tworzenia numerycznego modelu terenu (NMT) z danych lotniczego skaningu laserowego istnieje konieczność wydzielenia z surowej chmury punktów tylko tych,które były odbiciami wiązki lasera od powierzchni terenu. Zadanie to realizowane jest w znac
关键词：airborne laser scanning;filtering;flakes;variational problem
