Microtomographic Analysis of a Palaeolithic Wooden Point from the Ljubljanica River

Abstract

A rare and valuable Palaeolithic wooden point, presumably belonging to a hunting weapon, was found in the Ljubljanica River in Slovenia in 2008. In order to prevent complete decay, the waterlogged wooden artefact had to undergo conservation treatment, which usually involves some expected deformations of structure and shape. To investigate these changes, a series of surface-based 3D models of the artefact were created before, during and after the conservation process. Unfortunately, the surface-based 3D models were not sufficient to understand the internal processes inside the wooden artefact (cracks, cavities, fractures). Since some of the surface-based 3D models were taken with a microtomographic scanner, we decided to create a volumetric 3D model from the available 2D tomographic images. In order to have complete control and greater flexibility in creating the volumetric 3D model than is the case with commercial software, we decided to implement our own algorithm. In fact, two algorithms were implemented for the construction of surface-based 3D models and for the construction of volumetric 3D models, using (1) unsegmented 2D images CT and (2) segmented 2D images CT. The results were positive in comparison with commercial software and new information was obtained about the actual state and causes of the deformation of the artefact. Such models could be a valuable aid in the selection of appropriate conservation and restoration methods and techniques in cultural heritage research.

Keywords: 3D surface-based models; 3D volumetric models; Palaeolithic wooden point; archaeological documentation; computed tomography; computer vision; conservation; heritage science; waterlogged wood.

Figure 1

The 40,000 years old Palaeolithic wooden point, only one out of eight wooden artefacts of such age known in Europe, was found in 2008, submerged in the river Ljubljanica at Sinja Gorica [2]. As any waterlogged wooden artefact it had to undergo a conservation process to prevent it’s complete deterioration once taken out of water [4]. The wooden point before conservation (photo by Arhos d.o.o.).

Figure 2

Northeastern area of Vrhnika (Slovenia) with the Nauportus area (a; b) and the location (3) where the Paleolithic wooden point was found in the river Ljubljanica.

Figure 3

3D scanners used to capture the five 3D models of the palaeolithic wooden point in different years.

Figure 4

Index of change in volumetric data after the first scan (2009—index 100). The polar diagram shows the dynamics of volumetric changes (length, width, thickness and volume) between five 3D models of the Palaeolithic wooden point recorded between 2009 and 2018. The red line represents the volumetric state of the point at the first 3D scan (2009—index 100), which reflects the approximation of the state of the artefact under In situ conditions. The diagram clearly shows the changes that occurred at the beginning of the conservation process (2013—black line), when the artefact was subjected to intensive soaking and the addition of melamine resin (conservation). Thickness and volume increased significantly during this phase. After the drying process, the volumetric values (grey, yellow and blue line) decreased, especially the volume and thickness. However, slight dimensional changes in length and width were observed.

Figure 5

The deformation changes of the Palaeolithic wood point between the beginning and the end of preservation (volumetric changes, volume reduction and shape change) were calculated with the C2M (ICP) algorithm in the graphical software tool CloudCompare. The left image of the point shows the volumetric changes (red—bending of the upper and plant part; green—shrinking of the middle part) of the surface-based 3D model of 2017 compared to the reference 3D model (2013). The image on the right visualises a change in the shape of the 3D model of 2017 compared to the reference 3D model (2013). The colourimetric scale was created using an algorithm (C2M—CloudCompare) to statistically process the volumetric changes between the 2013 and 2017 models. The red-orange values represent the diffraction of the artefact from +3.6 mm to +1.2 mm. The blue-green values mark the shrinkage of the artefact between 10.4 mm and 1.9 mm. The data confirms the bending of the handle part and the top of the artefact. However, the shrinkage was more pronounced in the middle part. The cause of this deformation remains unclear. The unclear causes for the changes during the conservation process were the basic motivation for the creation of the anatomical 3D model of the Palaeolithic wooden point. By reconstructing the 3D model from 2D micro-CT images, we wanted to obtain volumetric data on the changes in the anatomical structure of the archaeological object.

Figure 6

Number of articles indexed in Google Scholar and published in MDPI journals during 2015-2021/11 that represent the use of computed tomography in the field of cultural heritage science and are directly or indirectly related to the topic of 3D rendering from 2D tomography images or 3D slices. No article was dedicated to the problem of direct reconstruction of 3D models from 2D tomography images. The 3D models were reconstructed using commercial industrial tomography software (VGStudio MAX [31,34,35,37], Amira Avizo 9.0 [32,33], Simpleware—Synopsys, Inc, Dragonfly Pro—Carl Zeiss) and in most cases analysed from 2D slices in different layers. In no case did we record the use of any of the open source programmes to represent a 3D anatomical model. Open source programmes do not currently provide data for advanced statistical and geometric analyses.

Figure 7

Computed tomography—presentation of the process of reconstruction of 2D CT images and 3D models.

Figure 8

Workflow of the direct dR3D and the segmentation algorithm sAR3D for reconstruction of 3D models from CT images.

Figure 9

Comparison of different edge detectors for segmentation of 2D slices CT. The figure shows the test results of the different edge-detection methods. The analysis and comparison of the test results was the basis for selecting the most suitable operator for performing the segmentation process in the phase of preparing the 2D micro-CT images for placement in 3D space and reconstruction of the 3D model. The Roberts Edge Operator was selected as the most suitable operator for performing the segmentation process. The advantage of the Roberts segmentation function is that it maintains the most small details without altering small cracks or indentations (these differences are more noticeable at higher magnifications).

Figure 10

3D model visualization process in the CloudCompare software tool. Read the explanations of the six steps in the text of the article.

Figure 11

Reconstruction results: (A) direct algorithm dAR3D, (B) segmentation algorithm sAR3D.

Figure 12

Example of a 3D model reconstruction with the algorithm dAR3D. The model detects anatomical changes (cracks, fractures, etc.), but the anatomical structure is also filled with woody parts. The reconstruction of the 3D model was performed considering all RGB values (0–255) of the grey matrix of the 2D images. Deformations of the internal structure of the model are only visible after individual sections. The model does not provide a detailed 2.5D insight into the artefact despite the large amount of information and data.

Figure 13

In the 2D CT image the edges of larger openings, fractures, pores, inclusions, etc., are detected first.

Figure 14

Palaeolithic wooden point: (A) 3D anatomical (volume) model, (B) 3D surface model, (C) 3D anatomical model with marked deformations (red dashed lines indicate the outer edges of cracks, openings and fractures in the internal structure of the point). The segmentation algorithm provides a 2.5D insight into the anatomical structure of the artefact after reconstruction. The outer surface boundaries of the artefact are marked in blue, with the light blue representing the inner openings, cracks and other deviations. The green colour represents the inner boundaries of the woody part of the artefact.

Figure 15

Exposed critical points in a 3D volumetric model of a Palaeolithic wooden point. A blue-green grid was chosen to make the anatomical structure of the artefact clearer. The light blue colour indicates the outer surface boundaries of the artefact and the inner boundaries of the non-wooden deformations (openings, fractures, cracks, pores, etc.) in the anatomical structure. The green colour indicates the inner boundaries of the wooden part of the artefact. The images show a view of the inner structure from the tip to the handle part (1’ and 2’) and from the handle part to the upper part of the artefact (1–4). Deviations and critical points are clearly visible.

Figure 16

Volumetric microlocations of critical sites in the volumetric 3D model of the Palaeolithic wooden point.

Figure 17

Overview of the critical points in the volumetric 3D model of the Palaeolithic wooden point.

Figure 18

Locations of exposed deformations of the Palaeolithic wooden point in the volumetric 3D model, which was recorded with a µCT scanner in 2019.

Figure 19

A three-dimensional depth image (viewed from the handle section) of the exposed critical areas in the anatomical structure of a Palaeolithic wooden point. Three main deformations were noted in the anatomical structure: a crack (B) running the entire length of the surface of the artefact; a transverse fracture (A) extending from the sampling point to the centre of the artefact; and numerous deformations, fractures, pores and cracks in the left wing of the artefact (C).

Figure 20

Fracture (A), which runs from the junction of the socket part and the point into the interior approx. 4.7 cm. A longer opening (B) is visible inside and cracks and fractures (C) in the left wing of the Palaeolithic wooden point.

Figure 21

Changes in surface deformation of the 2019 3D model of the Palaeolithic wooden point, compared to the 2018 reference model (changes are in a limited range between 0.0001 and 1.5001 mm). The colour matrix scale of deformation monitoring of the 2018 and 2019 3D models confirms the one-year dynamics of surface changes. Compared to other 3D models (2009–2017), the dynamics of changes on the artefact surface has stabilised. The shrinkage of the artefact persists with an average of 0.1 mm per year (green grid). However, the deformation of the uppermost point is even more pronounced. It lies between 0.5 mm and 1.0 mm (red-orange-yellow grid). The annual bending of the top by 1.35 mm is confirmed volumetrically. The bending is detected in the area of the handle. This has bent by 1.7 mm compared to 2018. More deformation of the left wing (A) of the point was also noted. In this area, the anatomical model drew attention to a number of unnatural internal cracks and deformations. Deformation variations in this area ranged from 1.1 mm to 2.2 mm compared to 2018, and the crack in the central part (B) widened by 1.4 mm. At the exposed point (B), it reaches a depth of 7.1 mm. Monitoring of the deformation was carried out using the CloudCompare software tool and the C2M algorithm (ICP).

Figure 22

Exposed sites of changes in the surface-based 3D model of the Palaeolithic wooden point (2019 —comparison with the 3D model from 2018). Volumetric measurements confirm the calming of the point deformation process. It is still dominated by shrinkage or bending in the range of 0.18—0.37 mm. Stand out (red value on the deformation scale—from 1.2 to 1.5 mm) deformation changes in the top, planting part and left wing point.

Figure 23

Annual dynamics of changes in anatomical structure (model—2019; reference model—2018)—view from the grip area to the top of the Palaeolithic wooden point. The deformation changes in the anatomical structure are clearly visible (crack along the entire length of the upper part of the artefact (1); larger fracture (2) in the lower and middle part; numerous unnatural deformations (3) in the left wing). The colour scale of the changes (red, green and orange grid) highlights the anatomical changes of the upper part of the artefact. The process of crack propagation and deformation of the tip is also shown.

Figure 24

Deformation monitoring of the top of the artefact (comparison of the 2019 model with the 2018 reference model). (A) shows the dynamics of volumetric changes on the surface, and (B) shows the dynamics of changes in the anatomical structure of the artefact. The colour scale represents the annual process of shrinkage and deformation of the upper part of the artefact. The dynamics range from 0.1 mm (green) to 1.1 mm (red). The upper side of the artefact is mainly exposed to a more intensive deformation process. On the inside, smaller cracks, openings and pores can be seen. These touch the beginning of the crack, which extends over the entire length from the top to the handle area. Due to the ongoing deformation process in this part of the artefact, the annual deflection of the top of the Palaeolithic head in 2019 was 1.35 mm.

Figure 25

Six examples of reconstruction of 3D models from CT images of various archaeological objects and other composite materials. On the left are three clay rattles [93], one the right a Neanderthal bone whistle [69] (top), a cylindrical piece of concrete (middle) and some fabric (bottom).