AMAD

„Archivum Medii Aevi Digitale - Interdisziplinäres Open-Access-Fachrepositorium und Wissenschaftsblog für Mittelalterforschung‟
 Zur Einreichung
AMAD BETA logo
Titel: Parsing Architecture within Plan Drawings with Application to Medieval Castles and Fortresses
Mitwirkende: The Pennsylvania State University CiteSeerX Archives
Autor*in: K. Galor
Y. Sui
A. Willis
S. Spina (editors
D. Pitzalis
C. Perlingieri
K. Debattista
Beschreibung: Plan drawings are graphical documents critical to the documentation of architectural features at historic sites. These drawings include important geometric information such as the location, shape, and size of architectural features, which, for decaying or collapsed structures, may be the only existing records of the intact structure. This paper discusses an algorithm that estimates the geometry and semantic interpretation of architectural structures from a plan drawing. The estimated values are used to automatically generate a 3D structure using the estimated semantic labels of structural elements in the plan drawing. We demonstrate the utility of this approach by parsing several plan drawings of medieval castles and fortresses and generating 3D reconstructions of these structures and detail typical circumstances that prevent the system from generating a valid reconstruction. Since the 3D model is derived from plan drawings where the architectural contour is well-defined, the approach automatically provides near-pixel level accuracy at all locations which is very difficult and time-consuming to guarantee when manu-ally constructing 3D models from the same drawing. Hence, these automatically-produced models can provide unprecedented accuracy to the in-situ remains not feasible with conventional manual model-building techniques. While this article represents initial work on this topic with limited scope (castles/fortresses), we envision that subsequent enhancements to this method will be a valuable tool for efficiently generating accurate 3D models for many different historic structures.
URI: https://www.amad.org/jspui/handle/123456789/73545
Quelle: http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.669.8868
http://visionlab.uncc.edu/downloads/arwillis/publications/2009_VAST_FloorPlanParsing.pdf
AMAD ID: 568405
Enthalten in den Sammlungen:BASE (Bielefeld Academic Search Engine)
General history of Europe


Dateien zu dieser Ressource:
Es gibt keine Dateien zu dieser Ressource.


Alle Ressourcen in diesem Repository sind urheberrechtlich geschützt, soweit nicht anderweitig angezeigt.