This discussion is intended as a brief introduction to the field of 3D digitization/documentation in Cultural Heritage. About its mission, means and future. Since this field is where I spent most of my focus, I thought it might be a good idea to draw some loose lines around what is what here.
What is 3D digitization
I wrote a few general things about digitization here and its distinct meaning from digitalization. I understand 3D digitization of the tangible cultural heritage as a process of transformation, or coding, the real-life characteristics (size measurements, geographical coordinates etc.) and chromatic aspects into digital 3D formats that can be interpreted and visualized with the help of computer software.
Scanning a document or a photograph for future digital usage is 2D digitization. For 3D digitization there are various methods to obtain a three dimensional digital replica of any kind of visible surface. The usual primary result in 3D digitization is the point cloud which is transformed into the 3D model.
As I’ll explain later, there are several methods for acquiring 3D data from physical objects. All these methods encodes in an internal coordinate system the shape and color of the digitized object. These shapes are collections of georeferenced data points (with 3D coordinates X, Y and Z) know as point clouds.
Point cloud generation is the most important step in the 3D reconstruction process. It is used for further processing for precise calculations or building 3D meshes (the popular 3D models). This format has the highest spatial resolution and is best suited for accurate analysis.
A 3D model is the digital representation of a shape or volume. People are usually refering to either the 3D object or the 3D file when using the term. 3D files contain the 3D object and are coded to be compatible for specific software applications. There is wide range of different 3D formats (files): OBJ, STL, PLY etc. Today’s 3D modeling, scanning and processing software can work with any of these formats. Which is great! Because not many years ago you would have to choose carefully the format you wanted to use in order to be compatible with a specific program.
3D models are either built from scratch by 3D modelling or generated from point clouds in the 3D reconstruction stages of 3D digitization.
Why is IT so important for Cultural Heritage?
In Cultural Heritage, as part of preventive conservation measures, 3D digitization has usually two main purposes: archiving/publishing and long term conservation state monitoring of the real surface.
While it would be worth a whole discussion to emphasize the importance of digitization in this field, for now I will just leave this comment by the Executive Director of Europeana Foundation:
Our museums, archives and libraries contain invaluable, often unique, records of heritage sites throughout their history – architectural drawings, photographs, paintings and written descriptions. Cultural heritage organisations are joining forces and adopting new technologies to preserve and share information about our heritage. By digitizing their valuable collections and making the data available to experts, they help to safeguard our heritage sites. In the face of today’s complex challenges, the task seems more urgent than ever.Harry Verwayen
But more on this topic in another discussion here.
The general pipeline of this process involves: 3D data capture, 3D reconstruction, geometric modeling and texture mapping. This will give you the raw model, the digital replica of the documented subject. Depending on the measurement’s purpose, from here on there are multiple workflows to get a final product ready for the next step.
Methods of 3D digitization
In Cultural Heritage 3D digitization, one of the important rules is: non-contact. The less physical interaction with the subject, the better. In this regard, I will focus here on the three main technologies that are currently used at global scale for this purpose. All these three technologies are using optical means for collecting data, therefore are safe for Cultural Heritage objects. These are: laser scanning, structured light scanning and photogrammetry. Newly emerged AI image-based technologies are still in experimental stages.
Automated Scanning systems
3D laser scanning is a well-established technique in industry but also in Cultural Heritage documentation field. Laser scanning devices are based either on time-of-flight method or on triangulation. Time-of-flight measures the time needed for a laser beam from its emission to the scanned surface and back to the capture sensor in order to calculate the 3D coordinate of the point of interaction on the surface. A system based on triangulation method projects a set of parallel laser lines onto the subject surface. A detector captures their reflection from a calibrated distance and calculates the 3D coordinates of the scanned points. These systems are fast and robust and can deliver results in spatial resolutions up to 0.02 mm.
Structured light scanning systems are based on the triangulation method, described above. Instead of laser beams, these devices are using white light or near infrared light. Like their laser-based counterparts, these are also robust and fast.
Both these technologies offer high spatial resolution and accuracy for the resulted 3D models and also color textures provided by embedded cameras. The commercial models of these systems are built and optimized for specific needs and casuistry: small objects, medium objects, large structures or aerial (LIDAR).
In one my first posts here, I wrote a lot about how photogrammetry is defined. If you want to learn about all the different definitions photogrammetry has or had during its long existence, and a bit about it’s history, come here. In this context it is enough to say that it is a passive optical method 3D documentation. It uses image measurement and interpretation for the determination of the shape and position of an object from one or more photographs.
At its core, photogrammetry uses the fundamental central projection model. Each photograph image generates a bundle of spatial beams, defined by the image points and the perspective center. If all the bundles from more images intersect, a dense web is created. Using the bundle adjustment method, a greater number of such bundles can be oriented. This way is calculated the three-dimensional coordinates associated to the object points. In order to apply these algorithms, the source images must contain the point we want to reconstruct in at least three different angles. This constraint results in a required partial overlapping of the source images.
Which one is better?
It is impossible to draw a clear line between these techniques and label which is better than the other or which is the best of them all. All of them are the best. But, in general, the best technique is the one that is available to you. If you have more than one available, use the one that best suits your time frame and job requirements.
With all these said let’s see what really differentiate these techniques. Photogrammetry is by far the most accessible (you really just need a decent camera and computing power, like all other techniques) and flexible (it can used in any case study). But for professional results it needs training in both photography and processing, and requires professional gear.
Laser and structured light scanning systems have the advantage of being on the market for a few decades now. They have matured and are currently used by all the professionals. Unlike photogrammetry that has re-emerged on the market a few years back (even though it is as old as photography itself) and, in my opinion, it still has a lot more to say in the near future (even though it has already matched and in some aspects even overtook the capabilities of the scanning systems). Scanning systems are fast, reliable and automated, but are optimized for specific situations (close up, close range, long range etc.). Texture on the other hand is limited and far inferior to photogrammetry.
To close this subject of comparison, let’s conclude with a few ideas. Scanning systems are reliable, fast and automated but they come with a high price (literally) and a mediocre color rendering. They are also constrained to specific situations (small objects vs building facades).
Photogrammetry CAN be free but it also requires an investment (but much-much less than with scanning devices), and the same equipment and software can be used for any situation (macro subjects, statues, buildings, aerial or even underwater!). It is best suited for polychromy and other situations where the real color recording is critical.
The downside is that this method is as reliable as the operator making the recordings. It directly depends on his/her skills for both recording but also for processing. It can be automated, especially for small objects. But in general, it comes with a whole legion of gear (lights, round tables, light modifiers, lenses, filters, tripod, light stands, backdrops, memory cards, external hard drives, lots of backup storage, image editing software, 3D reconstruction software etc.) versus a box rotating on a tripod linked to a laptop :).
The good news is that the high profile photogrammetry software is allowing the fusion of the results of these techniques into a complete new 3D model. Therefore you can benefit from both approaches given that you have available both methods.
Before you leave
This post is just a scratch on this topic. A really important discussion (which I’m gonna cover here) is the way 3D digitization is implemented in Cultural Heritage documentation. What risks or limitations can be met and especially what responsibility the operators must assume. Certain quality checks must be met otherwise we will be (and we already are) swimming in a sea of useless 3D models that cannot benefit in any way the Cultural Heritage.
If you’d like to refer to, or cite this text in your work please use the following resources as your citing source.
- Laurentiu Marian Angheluta, 2019. Practical Guide and Applications of Photogrammetry in Cultural Heritage 3D digitization, Editura Mega, ISBN 978-606-020-106-9
- Angheluță, L. M. and Rădvan, R.: Macro photogrammetry for the damage assessment of artwork painted surfaces, Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-2/W15, 101–107, https://doi.org/10.5194/isprs-archives-XLII-2-W15-101-2019, 2019
- [RO] Cercetări multidisciplinare pentru evaluarea stării de conservare a unor situri rupestre din zona Aluniș-Bozioru (Munții Buzău) incluse în patrimoniul cultural național (2017-2018)
- [RO] Practica digitizării 3D în documentarea continuă a Patrimoniului Cultural, MarketWatch.ro. May 2020, Nr. 224
Okay, that’s it for this week! Thank you for reading. I hope you learned a few things and, of course, for any questions please leave a comment below or shoot me a message with the contact form.
Have a great week! Cheers!