PREVIOUS   NEXT   CONTENTS   HOME 

3.2 Data acquisition strategies

There are many strategies for acquisition of these 3-D models (Böhler and Marbs 2002; Pavlidis et al. 2007; ter Haar et al. 2005). The most common are photogrammetric solutions and scanning (often laser scanning). Photogrammetric solutions use overlapping imagery and can develop highly precise 3-D models through the mathematical correlation of pixel locations. In laser scanning a laser beam is projected onto the object and the resulting location calculated. The resulting dataset is usually referred to as a point cloud.

When the projects were initiated in the summer of 2007 the Konica Minolta VIVID 9i laser scanner was chosen to scan and document both collections. The procedures used were essentially identical for both, though the field conditions at Amarna did constrain some options. The VIVID 9i is a close-range triangulation laser scanner that produces high-resolution, photorealistic 3-D models of objects. The scanner's technical specifications report an X, Y, Z accuracy of 0.05mm (Konica-Minolta 2011). While it may be possible that an individual scan may meet these specifications the process of scan merging and processing effectively degrades the results, as will be described below. We feel that the digital objects produced in these projects with this scanner can typically support metric analyses of properties that are in the range of 0.2mm or larger. A field-calibration system provided by Konica-Minolta is regularly used to calibrate the system and ensure that it is operating within the specified measurement tolerances.

With a set of three lenses the unit has a typical scan range of 0.5-1m. The system was used in conjunction with a calibrated turntable to help automate the scan process. We should note that since 2007 a number of newer close-range scanning systems have been developed that provide substantial improvements over the technical specifications of the Konica-Minolta system. Of particular technical note is that the Konica-Minolta uses an infrared laser system. This technology has many valuable uses but is not effective on a range of materials that are transparent, or semitransparent to the sensor (Hawkins et al. 2001; Secades 2008; Matusik et al. 2002). As a result materials such as feathers, fur, translucent stone such as some cherts, flint and marble, fresh bone and similar materials present challenges to the system. Additionally the system has a limited depth of field that created other challenges. Notwithstanding these technical limits the Vivid 9i served the project as a very effective tool. Higher precision instruments, such as the Breuckmann SmartScan HE that use structured white light, are now commercially available.

In the development of accurately textured 3-D models the capture of the range data and the capture of texture are related but somewhat separate; work flows providing different requirements and processing steps (Bernardini and Rushmeier 2002, 150). At the beginning of the project, it was clear that a professional lighting setup similar to that used for product photography was necessary to capture accurate RGB information for the artefacts. A number of previous authors have noted the process of merging multiple individual scans to create a complex 3-D object places substantial challenges on ensuring that the final colour textures have consistent and uniform lighting in both intensity and direction (Bernardini et al. 2001; Rushmeier and Bernardini 1999). Our approach was to light the object uniformly, using a professional lighting tent and professional lighting system with lights colour-balanced to the system's specifications. This minimised the effect of shadows across the object and reduced the impact of ambient light from the surrounding environment (Smallwood et al. 2009). The result is a more accurate capture of the true colour of an object's surface (see Fig. 1 for setup with VIVID 9i).

Figure 1

Figure 1: Scan setup with VIVID 9i and professional lighting rig.

At the beginning of the scan process, the turntable was calibrated with the system software, in this case, Polyworks (from Innovmetric Software), such that a single 360° rotation of an object, comprised of six scans, is automatically aligned. The object was placed on the turntable and a series of rotations conducted, governed by the size and complexity of the object. For a standard water bottle, a rotation is performed around the top and bottom of the object and then the object is laid on its side and a single shot is acquired of the base or underside of the object. In this case, a total of 13 scans were acquired. A scan typically consists of hundreds of thousands of individual points that have X, Y, Z surface information as well as RGB colour information.

The comments facility has now been turned off.


 PREVIOUS   NEXT   CONTENTS   HOME 

© Internet Archaeology/Author(s)
University of York legal statements | Terms and Conditions | File last updated: Tue Jun 28 2011