Hayat KADI | Karima ANOUCHE | Jean-Pierre PERRIN
(University of Sciences and Technology of Oran-MB, Oran, Algeria)

Keywords: Digital heritage, parametric modeling, 3D reconstruction, data acquisition

Within the framework of heritage documenting, three-dimensional reconstruction is the approach that currently offers the best compromise and protection against loss and/or oblivion. Although data acquisition has known significant progress thanks to the introduction of new tools such as 3D laser scanners, difficulties are encountered to effectively exploit and process the achieved data. Indeed, modeling element by element from the measured data is a time consuming. To assist clouds structuring, prior knowledge of the object may be the support of its digital reconstruction that requires geometric interpretations to generate consistent and re-usable models. From the observation and knowledge of the real object, it is possible to define a model describing the building, morphological relationships of its parts, and its geometric rules. It’s about describing all architectural components by a minimal set of parameters allowing their reconstruction.
This paper proposes a 3D reconstruction process that tries to facilitate transition from the acquisition and semantic description phases to a geometric modeling. Through the exploitation and digitization of the architectural and archaeological knowledge, this process aims to reach a geometric modeling, where parameterization serves as the principle of interpreting the acquired data. The examples of 3D reconstruction performed on columns and triumphal arches illustrate and explain the approach. It consists of extracting object components through depicting all existing typologies, and to identify parameters describing their geometry. The writing of corresponding algorithms then makes it possible to define parametric components to be implemented into parametric software. This allows the creation of new features and then generation of many models through multiples instantiations, which would adjust to available data (point clouds) for an accurate 3D reconstruction. Away from being solely a case study, the examples serves to explain our process meant to be extrapolated to other same-family objects and generalizable to external typologies, generating thereby models from a minimum of data avoiding the case-by-case basis.

Innovation: The 3D reconstruction process doesn’t intend to apply to limited case studies.