Using Semi-automatic 3D Scene Reconstruction to Create a Digital Medieval Charnel Chapel

Abstract

The use of a terrestrial laser scanner (TLS) has become a popular technique for the acquisition of 3D scenes in the fields of cultural heritage and archaeology. In this study, a semi-automatic reconstruction technique is presented to convert the point clouds that are produced, which often contain noise or are missing data, into a set of triangle meshes. The technique is applied to the reconstruction of a medieval charnel chapel. To reduce the computational complexity of reconstruction, the point cloud is first segmented into several components guided by the geometric structure of the scene. Landmarks are interactively marked on the point cloud and multiple cutting planes are created using the least squares method. Then, sampled point clouds for each component are meshed by ball-pivoting. In order to fill the large missing regions on the walls and ground plane, inserted triangle meshes are calculated on the basis of the convex hull of the projection points on the bounding plane. The iterative closest point (ICP) approach and local non-rigid registration methods are used to make the inserted triangle meshes and original model tightly match. Using these methods, we have reconstructed a digital model of the medieval charnel chapel, which not only serves to preserve a digital record of it, but also enables members of t he public to experience the space virtually.