Fully Automatic, Omnidirectional Acquisition of Geometry and Appearance in the Context of Cultural Heritage Preservation

Abstract

Effective documentation and display of ancient objects is an essential task in the field of cultural heritage conservation. Digitization plays an important role in the process of creating, preserving, and accessing objects in digital space. Up to the present day, industrial scanners are used for this task, which focus mainly on the detailed reconstruction of the object's geometry only. However, particularly important for a faithful digital presentation of the object is the appearance information-that is, a description of the used materials and how they interact with incident light. Using the world's first full-spherical scanner, we propose a user-friendly reconstruction process that is specifically tailored to the needs of digitizing and representing cultural heritage artifacts. More precisely, our hardware specifically addresses the problem that invaluable or fragile artifacts may not be turned over during acquisition. Nevertheless, we can digitize the object completely, including its bottom. Further, by integrating appearance information into our digitization, we achieve a far more faithful digital replica with a quality comparable to a real picture of the object. But in contrast to a static picture, our representation allows one to interactively change the viewing and lighting directions freely. In addition, the results are very memory efficient, consuming only several megabytes per scanned object. In cooperation with museums and a private collector, we digitized several cultural heritage artifacts to demonstrate the feasibility of the proposed process.

In search of the lost polychromy of English medieval alabaster panels in the Southwest of France

“…As in the previous works, this device (Figure 2A) consists of a robotic arm that moves a high‐speed camera (100 frames per second) and a “dome” containing 1080 LEDs, which allows sampling the light directions. The alabaster samples were disposed on a table that could be moved by a high‐precision rotation stage.…”

“…Their collective behavior is encoded in a 4D (two dimensions for the view direction, two for the light direction) reflectance function called the bidirectional reflectance distribution function (BRDF). 36 As in the previous works, [37][38][39] this device (Figure 2A) consists of a robotic arm that moves a high-speed camera (100 frames per second) and a "dome" containing 1080 LEDs, which allows sampling the light directions. The alabaster samples were disposed on a table that could be moved by a high-precision rotation stage.…”

3D mass digitization: a milestone for archeological documentation

“…While acquisition time is slightly higher the quality of the results has notably improved. At DFKI (Kohler, Noell, Reis, & Stricker, 2013) have built the ORCAM (Noell, Koehler, Reis, & Stricker, 2015), a fully spherical setup similar to the DOME which in addition to the previously mentioned setups is able to capture the bottom-side of artefacts by placing them on a transparent, rotational, anti-reflective glass carrier, pivo-mounted on a steel ring. Highresolution DSLR photo cameras and a structured light projector revolve around the sphere capturing geometry, texture and optical material properties.…”

Scene reassembly after multimodal digitization and pipeline evaluation using photorealistic rendering

“…Digitization is most often unimodal and tailored toward objects with a specific type of surface reflectance behavior [1]. While unimodal techniques are becoming more versatile [21][22][23], objects with a transparent material like glass still pose challenging problems. Their reflectance behavior is so different that they require an entirely different modality, such as computed tomography (CT) [24].…”