To preserve, study and present collections are three important principles of any museum. In maintaining these activities, museum institutions face a number of challenges, some old and some new. Firstly, after analysis, much archaeological material has to be repatriated to the country of origin. Secondly, technological development means museums are receiving more requests for destructive sampling. Thirdly, precious artefacts can be destroyed accidentally. Finally, expectations of museum visitors have increased and there is high demand for more information, and new and innovative media.

Sustainable solutions are needed for museums which face budget restrictions and which own extensive collections in need of digitisation. High resolution 3D digitisation can solve these problems by allowing museums:

• • to have continued access to specimens for scientific studies after destructive sampling or repatriation,

• • to create virtual 'back-ups' of objects as insurance against accidental loss,

• • to expand collections in terms of quantity and scientific value,

• • to create physical and virtual reproductions for museum display.

Figure 1.

Neanderthal talus of Spy (Spy Cave, Belgium; height = 6 cm) digitised with three techniques. RBINS Collections.

Over the last decade, 3D imaging has developed rapidly and several technologies are now available with differing resolutions, and also differing costs. The choice of technology depends on scientific purpose, but few comparative studies of the strengths and weaknesses of different packages are available (Breuckmann et al. 2009; Slizewski et al. 2010). The Belgian federal Agora3D Project (BELSPO AG/00/164) endeavours to develop guidelines in order to create 3D models with the best quality/price/time ratio for each type of natural history and cultural heritage collection. Agora3D is evaluating major digitisation technologies including computed tomography, µ-computed tomography, laser surface scanning, structured light surface scanning and photogrammetry. Each technique has advantages and drawbacks which the project aims to define.

To illustrate our purpose, we use a case study: the Neanderthal talus from Spy. The talus comes from a collection of Neanderthal bones belonging to two adults who were excavated at the end of the nineteenth century at Spy, Belgium. We have created several models of the talus using computed tomography, laser scanner and photogrammetry. All the examples seen in the figures here are also available in 3D PDF format which allow the reader to interact with and manipulate the virtual objects.

Figure 2.

3D model of a cutting head warrior from Costa Rica in basalt (height = 100cm), created by photogrammetry. In spite of the low relief, the engraving of the belt is recorded. RMAH Collections.

The first model was created by medical CT. Current medical CT technology is very fast and enables visualisation of internal structures, although the disadvantages are that it is necessary to take the specimen to the equipment and that the surface 3D model is computed from the threshold of the 2D X-ray data, which can produce size errors.

The second model was made with a low-cost laser scanner (NextEngine). This is a popular laser scanner used in many scientific institutions, which records both surface and texture. Compared with high-end equipment it is relatively slow, but the ratio between the cost and the result is highly favourable (Slizewski et al. 2010). The acquisition and the manual processing can take from several hours to several days depending on the complexity of the specimen.

The final model was created using photogrammetry which produces 3D models based on a set of 2D images. In the case of the talus, we used a 12-megapixel digital SLR camera with macro lens. The time needed for the recording (6 x 8 pictures) and the manual masking of the images was approximately one hour, followed by an automatic reconstruction of several hours. The model is scaled using either an internal scale or a measure made a posteriori on a set of landmarks. Photogrammetry allows the creation of precise models with rich texture using equipment and expertise which already exists in most museums. Moreover, the possible use of several lenses enables the digitisation of objects of different sizes (from millimetres to metres). The two examples displayed here show a very precise level of detail.

Figure 3.

3D models of a nucleus from Hélin quarry (Spiennes, Belgium; height = 7cm), created by photogrammetry. The flake scars and the riddles are recorded. RBINS Collections.

Photogrammetry is a long-established technique but recent developments in computer science have enabled models to attain excellent quality. We evaluated several packages and chose Agisoft Photoscan because it has a user-friendly interface and provides the best results in comparison to most of the other acquisition methods. The resolution achieved is better than medical CT and the texture has better colours than when using a surface scanner. The possibility offered by the program to remove surface texture also enables visualisation of micro-relief which is difficult to observe directly. Photogrammetry is also ideal for 'backing-up' objects before destructive sampling because the technique is non-destructive and has no effect on the artefact. Moulding or µ-CT acquisition are more precise, but they do not record texture; moreover, classical moulding is not always recommended for fragile specimens, and the impact of X-ray exposure on the different analyses has not yet been fully defined.

Finally, whilst digitisation with computed tomography and laser or structured light surface scanners require expensive equipment and a great deal of user experience, recent photogrammetry applications are much more user-friendly and have a less steep learning curve. The technique is also very useful in other contexts such as the recording of archaeological fieldwork (e.g. excavation, standing structures, rock art; Grussenmeyer et al. 2010; Reference ThanaphatarapornchaiThanaphatarapornchai, 2012).

3D digitisation has already proven to be a powerful tool both for the scientific study of museum collections, and for their conservation and display. The preliminary results of the Agora 3D Project indicate that photogrammetry can be a cost-effective technique for the digitisation of a large variety of artefact types, particularly in museums where the equipment and expertise required already often exist. The next phase of our research will explore other techniques such as structured light scanners with high resolution colour digitisation. At the conclusion of the project, we will create a summary of the advantages and disadvantages of each of the techniques tested. Among the aspects taken into account will be data quality, size of the 3D dataset, acquisition and post-treatment time, actual cost (of both equipment and staff) and the ratio between automated and manual work.