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：2006
卷号：XXXVI Part 5
出版社：Copernicus Publications
摘要：The paper deals with the terrestrial laser scanning survey of the Aquileia basilica, one of the most significant monuments of the Italian Romanic architecture, the religious centre of a small town in North-East Italy famous for its roman archaeological ruins. The laser data have been acquired with a Riegl LMS Z360i system integrated with a Nikon D100 metric camera, while the data processing has been carried out by the RiSCAN PRO software. The survey consists of 28 scans and 138 digital images for the internal aisles, and 14 scans and 55 digital images for the front external walls and the 73 m high bell tower. Thanks to 53 reflecting targets, surveyed from a GPS point network, the scans have been registered and geo-referenced in the Italian national cartographic system. An aerial laser survey of the whole Aquileia area, carried out with an Optech system, was yet available in the same datum. In this way, a very detailed 3D model has been reached for such a very complex architectonic monument: by suitably exploiting the software tools, a large number of numerical products have been automatically obtained, as coloured point clouds, TIN surfaces, vector sections, image texturing and orthophoto, as well as very realistic virtual navigations among them. For the automatic modeling of some main volumetric primitives, a three steps procedure, recently proposed by the authors, has been applied. By a nonparametric second order Taylor's expansion, the values of the 3D surface function, and the first and second order partial derivatives are locally estimated. From these, the elements of the so-called "Weingarten map" matrix are computed, and from the latter, the values of the gaussian and mean curvatures are finally evaluated. According to these values, the points are classified into four basic types (hyperbolic, parabolic, planar, and elliptic); after that, a region growing method allows a first raw segmentation. At last, a parametric regression for each raw cluster and a restricted surrounding buffer area is iteratively applied. The automatic modeling procedure has been applied to the basilica main apse. The obtained results seem to be very promising: the procedure has correctly classified the bottom parabolic part, the planes close to the windows and the top elliptical part
