Categories
3D-scan tutorial

Visualization and analysis of traces of use

3D capture of a Neolithic axe

Today I would like to talk about signs of wear and damage on archaeological finds. I have chosen a Neolithic polished rock axe as an example object. It comes from the central university collection of the Free University of Berlin and was originally picked up in preparation for my master’s thesis. Unfortunately, it is an old find with only moderate documentation. Therefore, although the find location has been handed down (Mielsdorf, Segeberg district), the circumstances of the find or its dating have not. Due to the non-glossy surface, stone tools like this are particularly suitable for photogrammetric methods.

To create the 3D model, 45 images were taken and cropped using Agisoft’s internal masking tool. The resulting model was calculated at high settings and originally had over 1,500,000 polygons. As it was still a very early model, a compact camera (Canon D10) was used as the camera. Unfortunately, the lack of manual aperture control led to some blurred areas in the texture.

Visualization of the 3D model in Blender 3D

The calculated 3D model is scaled to the original size of the axe and exported as an .OBJ file. This is then loaded into a prepared Blender file. I use various standard templates for render images in Blender, which are slightly different depending on the purpose of the render images.

Studio like render

The most common render images and animations that I create are calculated in a kind of virtual studio. In most cases, a soft, monochrome background serves as a base for the object to be rendered and any number of area lights as illumination. This quickly and easily turns a simple 3D model into a photorealistically simulated object that could also come from a professional photo studio. The paid Studio kit pro offers good support.

The finished animations provide a nice impression of the object and appear very realistic due to shadows and surface details.

Distortion-free recordings for publications

The second frequently used rendering variant is the distortion-free rendering of the 3D objects. I often use this template for boards or measurement images where aesthetics are less important than accuracy. The camera is switched from a standard perspective shot to orthogonal mode in order to create distortion-free views. The lamps are also not arranged as main light, side light and spots as before, but now simulate the most even illumination of the scene possible.

The biggest change is the background. I usually hide this completely, making it completely transparent. This not only massively reduces the calculation time of the images, but also automatically produces isolated objects. These can now be inserted into other scenes in an image editing program of your choice without having to laboriously isolate them. Unfortunately, the JPG file format does not support transparency. So I have uploaded the PNG files separately here and here.

The signs of use on the axe can be clearly seen in the images above. In the original color, they stand out from the polished green surface of the axe due to their darker color. In the color-correct image, the rougher flaking can be easily distinguished from the smooth surface.

Worn objects, “what’s missing?”

Signs of wear are varied and largely determined by the function of the object. In the case of this axe, it is mainly flaking that occurred during use of the object due to material fatigue. Larger defects can be seen in the area of ​​the blade in particular, as well as in the neck of the back of the axe, which may have been used as a hammer. But exactly how much material was lost due to flaking?

To find out, the 3D scan was converted from a polygon model to a voxel model. In contrast to polygons, voxels are objects with a volume and thus have mass. The gaps were then added manually in the 3D program (3D Coat) in order to obtain an approximation of the previously complete form. This reconstructed object can now be subtracted from the scan using a Boolean operation in order to obtain only the added areas. The new added areas are shown in red against the green original scan and can also be viewed live in 3D.

One of the greatest advantages of 3D models is the ease of determining volume. This means that highly accurate volumes can be derived quickly and easily even from the most complex shapes. Based on the original scan of the axe, a volume of 269.23 cm³ can be calculated for the object. The reconstructed chippings have a volume of around 15 cm³. Therefore, around 5.5% of the total volume was lost during the use of the tool before it was probably disposed of as no longer usable. In contrast, however, around 31% of the surface has damage.

It would be interesting to examine a number of axes from a similar time horizon to determine whether this represents a normal level of abrasion.

Volume and surface determinations can be carried out directly in Blender. To do this, install the freely available add-on 3D Printing Toolbox. Please do not forget to assign a scale and units of measurement to the scene. Otherwise, the output may be very abstract.


Are you interested in 3D scans, prints, reconstructions, workshops or questions?

Get in touch!

Leave a Reply

Your email address will not be published. Required fields are marked *