JOURNAL OF APPLIED ENGINEERING SCIENCES VOL. 2(15), issue 2_2012 ISSN 2247-3769 ISSN-L 2247-3769 (Print) / e-issn:2284-7197 3D SPATIAL DATA ACQUISITION AND MODELING OF ANGHEL SALIGNY MONUMENT USING TERRESTRIAL LASER SCANNING JOCEA Andreea Florina, Technical University of Civil Engineering of Bucharest, e-mail: ajocea@yahoo.com A B S T R A C T The 3D digital documentation of architectural objects and sites becomes more and more significant in the field of 3D modeling, where the visual quality and precise measuring have a great importance. In this paper are used the technological possibilities offered by measuring technique - terrestrial laser scanning and computer graphic software - for data acquisition as well as handling the acquired data. In the first part of the paper will be presented some basic principles and operational aspects involved in the application of terrestrial laser scanning technique. In the second part of the paper are presented the 3D modeling issues from which was generated a precise and accurate 3D model of Anghel Saligny monument placed in the courtyard of the Technical University of Civil Engineering from Bucharest, Romania. Received: July 24, 2012 Accepted: July 31, 2012 Revised: September 07, 2012 Available online: October 31, 2012 Keywords: measuring techniques, visual quality, modeling INTRODUCTION Over time surveying instruments has suffered a major transformation due to the progress in electronics and primarily due to the microchip development. The theodolite, the steel tape and the field book have been, in many cases, replaced by electronic field instrumentation such as EDM (Electronic Distance Measurement), GPS (Global Positioning System) and 3D laser scanning. Each technique came with something new: the EDM has introduced the electronically horizontal distance, GPS has introduced vector and at present 3D laser scanning, which is an advanced technology that is based on the latest laser techniques for distance measurement, generates a new set of information called as point cloud. While the laser technology for distance measurement is on the market for several years, the emergence of terrestrial 3D laser scanning systems grew over the last years. Until now, a classification of these systems has been realized taking into account the distance measuring principle, in geodetic applications imposing the continuous wave principle and phase difference principle. There are, also, special scanners for near field (up to 2 m) which are based on triangulation principle. Also, in order to compare different products of terrestrial 3D laser scanning it should be considered some aspects related to technical specifications such as resolution expressed as number of points/steradian, spot size of laser beam and scan rate expressed in measurement number/second. Terrestrial laser scanning is a technique which allows the determination of geometry of a structure completely automated (more or less) without a reflective environment, with high accuracies and high speed. The space object is swept by laser beam on columns or lines, the resulting points forming in their entirety a so-called point cloud [Figure 1]. Terrestrial laser scanners are sensors that allow the registration of 3D position of objects by measuring a horizontal and vertical angle and a spatial distance to each point. By using trigonometric functions can be obtained the points coordinates in an internal coordinate system of laser scanner. These coordinates can be transformed by georeferencing in a proper (X, Y, Z) coordinate system.
Fig. 1. Terrestrial laser scanning principle [4] The advantages of this survey procedure are the high speed and the high point density. Due to these abilities it is possible to be applied in various fields such as: archaeology, architecture and cultural heritage. Taking into account what was have mentioned above, terrestrial laser scanning is a suitable and efficient technique for recording geometrical 3D object recording of complex objects such as monuments. In this paper we it will be presented the geometrical 3D object recording and modeling of Anghel Saligny monument placed in the courtyard of Technical University of Civil Engineering of Bucharest, Romania [Figure 2]. The 3D spatial data acquired using Leica ScanStation 2 Terrestrial Laser Scanner was processed using Cyclone and Geomagic Studio software. Main issues pursued in this application were the degree of automation in data acquisition, data processing and the final result object modeling. Fig. 2. Anghel Saligny monument location MATERIALS AND METHODS 1. Data acquisition Before starting the scanning process, three reflective targets were placed near the area of interest, Anghel Saligny monument [Figure 3].
JOURNAL OF APPLIED ENGINEERING SCIENCES VOL. 2(15), issue 2_2012 ISSN 2247-3769 ISSN-L 2247-3769 (Print) / e-issn:2284-7197 Fig. 3. The area of interest and the reflective targets Data acquisition was carried out in a local scanner system. The entire structure was scanned from 3 scan stations in one hour and 45 minutes. The scanning was performed with an accuracy of 3 mm. A total of 5.5 million scanned points were stored in a file of 184 Mb [Figure 4]. Fig. 4. The area of interest and data acquiring scan stations 2. Data processing The data processing was carried out in the first step with Cyclone software, then with Geomagic Studio software. The first step in data processing was the transformation of all point clouds from the local scanner system into a common system (so-called registration process). This step was achieved in Cyclone software using a 3D Helmert transformation (7 parameters) without taking into account the scale factor. The scanning and registration operations were performed automatically using reflective targets. The targets were recognized semi-automatically by the program during the scanning process. The registration process was realized without inconveniences due to the fact that we had only 3 scan stations and the targets were correctly acquired during scanning process. Thereby a precision of 1 mm was obtained [Figure 5]. Fig. 5. Registration process result
The second step of processing was the generation of the virtual 3D model. For this the registered point cloud was filtered and consequently the volume of data was reduced [Figure 6]. a. b. Fig. 6. a.the registered point cloud; b. The filtered point cloud 3. Modeling The raw data supplied by the laser scanner systems are not always well suitable for a direct use in CAD systems. After the cleaning of the 3D point set, a simple and more direct way to produce a surface, that can be sufficient for some applications, is to generate a polygon mesh. A polygonal mesh is described by topology and geometry. The topology of the mesh is given by the neighborhood structure and the geometry is defined by the coordinates of the vertices. The topological structure of the mesh contains vertices, edges, triangles and other features like holes, genus and number of connected components. The relation between these elements is given by the Euler-Poincare formula: V-E+F-H=2(C-G) (1) where V is the number of vertices, E the number of edges, F the number of triangles, H the number of holes, C the number of connected components and G the sum of the geni of all components [2]. The simplest type of polygon meshes, currently used in most of the commercial scanning or reverse engineering software, is the triangular mesh. This type of mesh is usually generated using the Delaunay triangulation method [5]. The geometrical dual of the Delaunay triangulation is the Voronoi diagram [Figure 7]. It essentially divides the space into convex cells by creating a region around each known site pi such that every point from the region is closer to the point p than to any from the other sites [5]. The topological elements of the Voronoi diagram are: the Voronoi point (the point located at equidistant distance from three different sites), the Voronoi line (given by the points which are equidistant to two sites in the plane) and the Voronoi cell or polygon [1].
JOURNAL OF APPLIED ENGINEERING SCIENCES VOL. 2(15), issue 2_2012 ISSN 2247-3769 ISSN-L 2247-3769 (Print) / e-issn:2284-7197 Fig. 7. Delanay triangulation vs. Voronoi diagram [1] In order to process the polygonal mesh by producing the final model of Anghel Saligny monument, a set of procedures should be performed to detect and fix topological or geometrical defects, to remove unwanted or unnecessary data and to create a complete mesh [Figure 8], [1]. These operations were carried out automatically in Geomagic Studio software. Although the statue case is not very complicated to carry out the task, in the present case study were used almost all the above software applications. Fig. 8. Topological and geometrical errors After performing the automatic computation of the triangle meshing, the result was optimized afterwards, the existing holes were filled and the surfaces were smoothed. The final result is presented in Figure 9. Fig. 9. The optimized result During the smoothing of the surfaces one had to proceed very carefully in Geomagic Studio, since important details and intricacies of the surface are quickly lost by overregulation of the parameter settings. Generally while handling this dataset with Geomagic Studio it was revealed that experience was necessary for optimal parameter control [3]. A geometrically correct 3D model of
the monument [Figure 10] has been generated by processing of the complex object areas (free form surfaces) with Geomagic Studio, which demonstrates a very high visual recognition value [3]. Fig. 10. The final result Anghel Saligny Monument CONCLUSIONS This paper describes the use of terrestrial laser scanning as main procedure of data acquisition in field of cultural heritage. The practical example demonstrates that the tested terrestrial laser scanner system (Leica ScanStation 2) is suitable for detailed data acquisition and object modeling providing high qualitative data. REFERENCES 1. DUESCU E. (2006), Digital 3D documentation of cultural heritage sites based on terrestrial laser scanning, PhD Thesis, Germany, pp. 58-59. 2. KARBACHER S. & CAMPAGNA S. (2000), Principles of 3D Image Analysis and Synthesis, B.Girod, G. Greiner und H. Niemann (Ed.), Kluwere Academic Publishers, Boston Dordrecht-London, pp. 141-143. 3. KERSTEN T.P. (2010), 3D scanning and modeling of the Bismarck monument by terrestrial laser scanning for integration into a 3D city modeling of Hamburg, M. Ioanides (Ed.):Euromed 2010, LNCS 6436, pp. 179-192, 2010, Springer-Verlag Berlin Heidelberg. 4. NEUNER J. (2009), Sensors Course Note, Technical University of Civil Engineering of Bucharest, Faculty of Geodesy, Romania. 5. PREAPARATA F.P & SHAMOS M. I. (1985), Computational geometry: an introduction. Springer-Verlag.