Fabrizio GALEAZZI / Stefan LINDGREN
(University of California, Merced – Lund University, USA)
Keywords: 3D documentation, archaeological excavation, laser scanner, dense stereo matching, accuracy.
Abstract:
The use of three dimensional technologies for the documentation of archaeological and cultural heritage sites is well established today, but this can be challenging because researchers have yet to integrate these technologies to develop a complete and coherent methodology for the 3D documentation of sites. Laser scanning and, recently, dense stereo matching techniques have shown to be very powerful tools for the 3D documentation of the archaeological excavation and context.
The proposed work tests and compares laser scanning and dense stereo matching techniques at Las Cuevas site (Belize), to find the most appropriate methods to document different aspects of an archeological site, based on diverse environmental conditions, light exposures, and varied surfaces. The complexity of the Las Cuevas’ site allowed testing the 3D documentation techniques in areas of the site presenting different environmental conditions and light exposure: dark recesses of caves; areas in shaded sunlight under the jungle canopy; and places of more direct sunlight in areas that have been cleared of brush or exposed by tree fall.
This paper presents the results of the metrical comparison between the 3D models obtained using phase shift variation laser scanner (Faro Focus 3D) and dense stereo matching software (Agisoft Photoscan). The comparison of the co-registered surfaces was performed in both open source and commercial point cloud and mesh processing software based on the shortest point-to-mesh distance considering the normal to the mesh faces.
This study points to dense stereo matching as the most economical, portable, flexible, and widely used approach for the 3D documentation of archaeological sites today. Nonetheless, the result of the quantitative comparison underscores the need to integrate this technique with laser scanner technologies, when the data acquisition of micro-stratigraphy is required. This kind of research can help defining best 3D practices for the documentation, visualization and analysis of sites.