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：2009
卷号：XXXVIII-5/W1
出版社：Copernicus Publications
摘要：Vehicle based mobile mapping technology has been developed quickly in the last decade. The high productivity is the main advantage of mobile mapping systems. Hundreds of kilometers are achieved by mobile laser scanning and the raw point clouds are almost saved without further treatment. It could be noticed that the automatic treatment of these clouds has not developed as quickly as the acquisition itself; hence the data are accumulated until an efficient method of treatment is available. After a brief presentation of the low cost mobile mapping system, this paper shows the first results obtained by application of a contour extraction algorithm on provided data. A prototype of low cost mobile mapping system has been developed in the MAP -PAGE laboratory for research purposes rather than production ones. The relative accuracy is provided by the system for every point, as auxiliary data. Furthermore, the geometry of mobile mapping point clouds is different from that obtained by a fixe TLS. Such particularities should be taken into consideration while processing this type of point clouds. Thus, the segmentation and edge detection approaches, intended initially to treat stationary TLS data, have been adapted to the particularities of mobile one. The segmentation is based on hypothesis that the most prominent features of fa.ade components are planar. Therefore, a segmentation based on RANSAC algorithm has been applied in a sequential way in order to extract all potential planes from the mobile point cloud. Corrective operations aiming to define the most reliable points of each plane have been implemented. A second algorithm aims to extract the boundary points of the planes resulting from the previous segmentation. This extraction algorithm relies on a Delaunay triangulation based on the lengths of the triangle sides composing the whole triangulated network. This algorithm is able to extract automatically and simultaneously not only the outer (wall borders) but also the inner contours (holes like windows) of fa.ade elements. Finally, in order to validate the results, large samples with different characteristics have been acquired by the mobile system and have been introduced in the developed workflow. The limitations as well as the capabilities of each process will be emphasized. Results are satisfactory and confirm that the acquisition as well as the processing approaches are reliable for generating a vector model of building fa.ades
关键词：TLS; GPS; IMU; P oint cloud; Mobile mapping; Algorithm; Building; Fa.ade; Extraction
