Dr. Smiti Nathan

Transcript

Smiti: Welcome to Curiosity Meets the Past, a podcast for curious minds who love learning about the past. My name is Dr. Smiti Nathan, and I’m an archaeologist and your host.

In this episode, I’m speaking with archeologist Madeline Robinson, who’s currently a PhD candidate at the University of Sydney in Australia. Her research utilizes various digital methods to better understand the past. These include photogrammetry, GIS, remote sensing, and more.

Her research utilizes various digital methods to better understand the past. These include photogrammetry, GIS, and remote sensing. Now these techniques can be used in a variety of ways, and Madeline’s specialty is using them for 3D modeling.

In our conversation, we discussed some of the ancient sites and artifacts that she’s researched using 3D modeling and what these models have revealed about the past that we otherwise might have not known.

She also shares her biggest surprise so far with a 3D model, and her story got me really excited about this application and how it could potentially help people engage with the past.

She starts us off by sharing the first time she encountered one of the methods that she now heavily uses to make 3D models.

Madeline Robinson: So the first time I ever heard of photogrammetry was about 10 years ago during a maritime archeology course run through Flinders University at Mount Dutton Bay in South Australia. And one of our tasks that we had to do was to learn how to model a shipwreck. Which was actually a type of oyster cutter.

It was only about 12 feet long, not massive. but we had to model it whilst scuba diving. and this is when Dr. Kōtarō Yamafune, was one of the staff members running the course, literally changed the course of my life by introducing a type of 3D modeling called photogrammetry. And what really struck me, so my aha moment was that this method. Brought together two things. I already loved photography and archeology, so I’ve been an amateur photographer since my teens. And suddenly these skills weren’t just creative. They were a way of digitally preserving archeology, it went beyond preservation. So once we had dredged this wreck and revealed its full shape and morphology, we were able to record it quite quickly and create a 3D model that could be analyzed again and again without tanks on our back. Without time limits and without physically impacting the site. knowing the wreck would continue to grade, especially ’cause we had dredged it. and then, you know, as soon as we had reported it, it was gonna upset ’em. It’s gonna fill in again, capturing its condition at that moment. Time was incredibly powerful, so that experience completely changed how I saw archeology and opened up a whole new direction of archeological practice that I didn’t know, I didn’t realize existed.

Smiti: That’s really amazing and I love hearing that story. Now, there’s a lot of different types of 3D modeling techniques and I was wondering if you could give an overview of photogrammetry, for our listeners who might not be as familiar with the term or that type of 3D modeling.

Madeline Robinson: Photogrammetry is a way of creating accurate 3D models using overlapping 2D photographs. So by taking lots of images, using a mirrorless or DSLR camera from different angles and tip different perspectives, we can upload these images into modeling software that works out depth, shape, and scale. A bit like how our eyes create, depth perception and uses those images to build a color meshed and textured 3D model. 

So for archeologists, this means we can digitally capture artifacts, sites, and even whole landscapes, both above and underwater in incredible detail. So those models become long lasting digital records that we can measure, study, and revisit long after. We no longer have physical access to the site or the object.

Smiti: Thank you for that overview. So everyone is up to speed and you’ve done photogrammetry in a number of locations. and one of the places that I saw you’ve done work in is a place we’ve mentioned on this podcast before, which is Malta. The Maltese megalithic monuments also called the Maltese Temples. I was wondering if you could share a bit of that experience with us and what you were doing there.

Madeline Robinson: So making the models of the Maltese temples was incredibly fruitful. What the photogrammetry models revealed was the underlying engineering of the structures, things that aren’t immediately visible when you’re standing in front of them and looking at them today. So as we all probably are aware, only the walls and foundations remain.

The roofs are completely missing on all the temples, raising a lot of conjecture about, you know, what they look like and how they were made. so by creating detailed photogrammetry models of, I did three temple complexes, so Mnajdra, Tarxien, and Ġgantija, I was able to strip away all that surface texture. So I took away the color, the decoration, all of that.

I pulled that off and looked purely at the morphology of the structures. So once I did that, architectural details started to emerge that were not obvious in the field. I could see groove marks that pointed to earlier construction, and engineering features that were just no longer there. So that opened the door to testing reconstruction ideas.

And the way I did this was that I compared my temple models with the  available archeological evidence for the original architecture. So there is actually quite a lot of iconography, that shows the original structure, including clay models. We have a carved limestone little block of, one set of it’s one set of apses, it’s like an oval shape and it shows the complete, temple structure. But the most important reference came from the Hypogeum of Ħal-Saflieni. And that is a three layered, 10 meter deep, burial chamber. And so whe n they first found it, it was filled with skeletons. 

So it’s located right next to the Tarxien Temple Complex. So they’re obviously associated also in the same time period. And in one of those chambers is, a room called the Oracle room. And the Temple Builders actually carved the architectural design of the temples directly into the rock. And the crucial thing here, here is that they carved the same trilithon structure that we see in mostly all the temples across Malta.

So this trilithon structure is. It’s very recognizable and what it is, is that you’ve got two upright orthos stats or standing stones with a lintel piece that sort of. Found the portal, the movement between spaces, if you come from like outside into the temple, and then further into the next set of apses, these trilithon structures sort of signpost that movement.

It’s like a very grand architectural signpost for liminal movement. And so outside of this portal, you’ve got two false windows and then another set of Stats these upright limestone blocks like this. And then you have a big lintel piece that sits on top. So they had carved that triathlon structure into this oracle room, and then they continued what the ceiling would’ve looked like.

So this was key. and very fortunately, Malta heritage who run and maintain the temples, they created a LIDAR model, of the Hypogeum and they gave it to me, which was amazing. So what I did was I cut out, I sort of extracted the oracle room and with very little resizing, repositioned that onto my photogrammetry models.

So that allowed me to test was this feasible? Did it fit, did it work? Was it actually architecturally, sound? was it structured in a way that could support all that, you know, massive, either limestone or timber? We don’t know which it would’ve likely been, but I did test both scenarios and yes, they were both feasible, but the crucial thing that was so important was that.

Getting the, getting that LIDAR model positioning onto my models, my temple models, the grooves were consistent. So where that slotted in was where I could see on the photogrammetry models where those original groove marks were. So this was such a big moment, like there was no other way I would’ve been able to do this and verify with the evidence that this was feasible.

I just wouldn’t have seen those grooves  and. I didn’t see them in the field even when I was like staring at these temples for weeks and either with my own eyes or through a lens, I still didn’t see them. But additionally, despite only being in Malta for a few weeks, those photogrammetry models allowed me to study the engineering of the temples in detail for well over a year.

And even today, I occasionally still look at them. You know, this is, this is a snapshot of their, their condition in time that. I can look at whenever.

Smiti: That’s so powerful to be able to reference those images time and time again, especially when you’re not in the field. But what I also really appreciate about that story is how you use multiple methods together to, figure out something or an aspect of something in this case, like Maltese construction techniques, which is a huge contribution.

And we’ll make sure to link the, peer reviewed article, with, Madeline’s results, below. But now, I would like to transition to talk about another use case for photogrammetry, which is artifacts. And you’ve done this type of 3D modeling on a specific artifact class, and I was wondering if you could share a bit more about that.

Madeline Robinson: So I worked for many years as a research assistant developing photogrammetry methods. Specifically for stone tools, and they ranged, some were really beautiful large hand axes, but a lot of them, and the majority of them were microliths and micro liths are really teeny tiny stone tools, and the ones that I had to work with were more or less under two centimeters long.

They’re really small. Often really irregularly shaped, and so difficult to model. They required so much more like higher precision  , photogrammetry method and equipment than standard object photo, um, photogrammetry also, I had so many to work through. So I needed a method that was consistent and repeatable.

and that became just as important as model quality. So developing the capture workflow, it did take a long time because these objects were so small and so fiddly and there wasn’t a plug and play process. And it involved extensive testing of shooting environments. So I tested in a light box, In using turntables and different arrangements of, different types of, of light boxes actually as well. different camera setups, so different positions capturing different angles, you know, the number of rounds, per angle. And I tested so many different macro lenses. I think I tested six different macro lenses ranging from, a focal length of like 35 mil.

To, I think about 120 or 150. And I even played with a 600 mil macro lens, which was way too difficult to work with. but then I had to play with the lighting, so the direction, intensity, color, temperature, and their position. And then I had to think about the object orientation. and also it’s stability because they’re so small.

A single breath, make them move. And I needed the shots to be really crisp and clean, you know, so it was really difficult. because microliths, they, they vary so massively, even though they more or less just made of church blin. they varied so much in color in translucency and reflectiveness, which creating photogrammetry models is really difficult if something’s translucent or is very reflective because some absorb light and others scatter it. So it was genuinely really frustrating. I had lots of trial and error and a lot of reshooting because the goal here, wasn’t just to visualize, the models. The aim was to produce accurate analytical 3D models that could be used to assess manufacture techniques, examine reduction sequences, and analyze flake scars and edge modification.

Through extensive trial and error, I developed multiple capture approaches, and this was dependent on size, shape, material, properties. So at the end of it, there was no one size fits all solution. It was like, okay, depending on what variables. This is the method I’m gonna follow. So this was a huge turning point when I realized this.

And once the methods worked, they worked reliably. So it enabled large scale repeatable modeling of micro liths, and it showed that photogrammetry could be used for serious lithic analysis, but it also expanded where object-based photogrammetry can realistically be acquired.

Smiti: I can completely empathize with trying to analyze a certain type of artifact and it feeling very tedious and hoping to create some type of semblance of a workflow. So that’s great that you were able to create something, specifically a workflow that worked for, these stone tools that you were looking at. I’m curious with all your experiences in creating 3D models for landscapes and artifacts, has there been something that’s surprised you as a 3D model was, forming on your end?

Madeline Robinson: So I think all photogrammetrists can agree. Anytime you see a model come together, you do get a, a real kick like of satisfaction and it’s so lovely to see things come together, especially when you’ve been working so hard. On, you know, pulling these models together. Some are really, really difficult. But I think, my biggest surprise moment was nothing that I saw on screen, it wasn’t the first time that I got a 3D model printed.

The, the biggest surprise by far was the first time I was able to walk through my models on virtual reality. I had the models projected to scale. And as soon as I put the goggles on, I was no longer in our computer room. I was in the temples again, and it was so surreal. I nearly broke my leg running to like run around the temples.

Um, nearly damaged so many desks on computers because. It was so real. It felt like I was back there again. I don’t work a lot with vr. I do the, you know, the analysis and the, like, the teaching photogrammetry, but the, the experience and people who have done VR will know exactly what I’m talking about.

I was walking on one of the walls at, I think it was the Southern temple of Mnajdra. It’s got a really, really tall wall. It’s about five and a half or six meters, and I’m on top of the wall ’cause I somehow position myself up there and looking down, getting such like bad vertigo. ’cause I get vertigo anyway.

And my God, that just, the, the feeling like your brain is telling you you’re gonna fall, you know, you are, you are, you’re believing what your eyes are seeing, which is, you know how our brains work. So it was really, it struck me how immersive it was. It was so eerie, almost like being back in Malta again. 

But beyond my personal reaction, Something became very clear. It meant that I could share this experience about being in the temples with other people. People who may never travel to Malta or see the temples. They could still experience their scale, you know, their grandeur, their spatial presence, not as a replacement for being there, but as a meaningful way of engaging with them.

And that’s really one core powerful thing about photogrammetry. It’s accessible. It brings people together and opens up opportunities for people to experience places and heritage that might never otherwise have have access to. So I think it’s a really, it’s a really beautiful and amazing thing.

Smiti: That was such a vivid example, and I appreciate you sharing that and discussing the accessibility aspects of photogrammetry. So along those same lines, you wrote a book about photogrammetry and archeology, and one of the things I personally love about the book is that it feels that anyone who is interested in learning about photogrammetry and how to actually do it could read your book and follow along, and learn how to do it. I found the book to be very accessible in that regard. So I was wondering if you could share a bit about your thought process in creating and writing the book, and what your goal was with it.

Madeline Robinson: So the book really grew out of my own research and teaching as I was developing those photogrammetry methods, as an RA, I started writing down my workflows because I have a terrible memory and I need to write everything down. But then, then suddenly realized I had a lot of material. And at the same time, I realized just from talking to a few students at, the University of Sydney, there were many of them that were desperately wanting to learn how to use photogrammetry and apply to their own research projects.

So a friend of, mine and I, we were like. Let’s start a, a group. Let’s start a, a photogrammetry student club at the Uni University. and we called it Three dig, which I loved. It was 3D Imaging Group. I thought it was very well-known. and I began teaching and refining these workflows that I’d written down with students.

So that experience was crucial because I didn’t just learn how to explain photogrammetry. I learned how students actually experience learning it. So I learned how to teach them in a way that was successful, that they responded to. 

And what became very obvious, very quickly was that most students don’t struggle with the idea of 3D modeling itself. They love that part. Seeing a model come to life is just so exciting and, and so rewarding. 

But where they got overwhelmed, Is much earlier at the stage where you have to set up and configure the camera setting to obtain correct exposure. It was the principles of exposure that would freak people out and they would get really, really overwhelmed.

And it was that point where they realized, you know, hey, on photogram, I heard photogrammetry was really easy. You just point and shoot and you, you thro w into a software and there’s the model. no, it can be really technical and that’s started to stress out a lot of students, so. Because of that experience and knowing how people responded to this stuff that I had taken for granted since I was teen.

I know how to set up a camera. I’ve been doing it for like 20 years, but it was scary for people who didn’t know. So I approached writing the book. In a very intentional and gentle way. I didn’t want to scare anyone. I wanted to introduce topics like, you know, terminology, equipment, and concepts. I introduced ’em slowly, and in very simple terms, which actually is really quite hard to write things like complicated things quite simply.

But I wanted to help readers understand. Why the things work the way they do, especially in the image capture stage, which is the most critical part of the entire process because if you take poor photographs, you’ll be left with a poor model and it will take so much more time and energy to fix it and post-processing, to pull or, you know, all the things together to make it work.

So the goal was to take, what can seem like a complex or specialist knowledge and break it down into principles that anyone can understand. Because you don’t need to be a photographer to do photogrammetry. You really don’t. You just need to understand a few core ideas. So ultimately, I wrote the book for students who are curious and motivated about photogrammetry, but maybe be a bit intimidated because the way that I’ve written this, you don’t need any prior knowledge.

You really don’t. I just walk you through each step along the way. and honestly, if the book helps, even just one student feel confident enough to complete their first modeling project, that’s a huge win for me.

Smiti: I am sure it will, and it’s already helped me in terms of thinking how to frame certain aspects of digital methods that we use in archeology. So again, I think it’s a fantastic. Book So technologies can often change quickly, and I’ve seen a lot of technologies change and adapt or become obsolete in the last 20 some years of doing archeology, especially when it comes to digital methods. So I’m wondering what do you think is the future of photogrammetry?

Madeline Robinson: I firmly believe that the form of photogrammetry that we’ve all been working with, what I’ve been working with, over the past decade is an essential foundation that we had to explore as a steppings. stepping stone for whatever is coming next. I have no doubt that many of our current methods will eventually become obsolete.

They have to, they always do. 10 years ago photogrammetry felt novel and experimental and exciting, and now it’s, it’s commonplace. And the same thing, thing goes for LIDAR really, the idea of having it, you know, on a phone would’ve seemed unbelievable not that long ago, and now it’s everywhere. I can generate a measurable 3D model, on my phone in minutes.

The pace of technology change. Is incredibly fast and we will absolutely see newer, faster, and more automated methods emerge. I do reckon, reckon laser scanning will likely continue to overtake passive remote sensing techniques like photogrammetry. in some context. simply because it’s quicker and less dependent on lighting or surface conditions.

but that doesn’t make photogrammetry irrelevant. It was a crucial step in getting us to where we are now and where we’re going next. And without it, we would not have developed the workflows, the critical thinking, or even the expectations that we now have about, uh, 3D data. So people sometimes ask me, you know, why would I write a book about a technology that’s constantly changing?

And honestly, it doesn’t worry me at all, my book and my contributions more broadly. They’re just part of the record of how archeology learned to think and record in three dimensions. So just like photographs, 3D models will remain valuable. They’re visual records that preserve archeological information, and even as technologies evolve, those records don’t disappear.

They just become part of the story.

Smiti: That was a beautiful answer. I’m also curious to see the evolution of photogrammetry as well as how it’s used with some of the other methods you’ve mentioned, like lidar. And as we wrap up here, what do you hope listeners take away about modeling can reveal about the past?

Madeline Robinson: Okay, so there are three things I really hope listeners take away from this. The first is that, we often underestimate the power of 3D models. They’re sometimes dismissed as like inferior replicas or as gimmicky tools used mainly for engagement rather than serious scholarship.

But I see them very differently. I see them. As like powerful surrogates, they allow us to work with archeological material without physically impacting it. And in a discipline where so much of what we do is inherently destructive, that matters enormously. the second thing is that 3D models act as a record in time, and this is where their analytical power really comes in.

So archeology is largely irreversible. Once something is excavated, it can never be put back. Exactly. As it was by creating a 3D model, we capture an object or site at a specific moment. Its condition, its form, its volume and measurements. And that record allows us to slow time. It lets us revisit material, reanalyze it, and ask new kinds of questions long after the physical context has changed or disappeared.

So when you can measure, compare, and visualize things in three dimensions, patterns and relationships emerge that simply aren’t obvious in two dimensions or in the field. And the third and final thing that I would like people to take away from today’s conversation is that. 3D modeling fundamentally changes who gets access to the past, and this is really crucial.

So instead of knowledge being locked away in, you know, storage rooms or labs, or behind paywalls, these models can be shared with students, communities, and the public anywhere in the world. And this has real implications, not just for education, but for equity inclusion, and even repatriation. It doesn’t replace traditional archeology.

It just reshapes who gets to engage with it and how.

Smiti: One of my biggest takeaways from this conversation is that when we approach digital methods like photogrammetry through lenses of care, education, and access, we actually can engage a wider range of folks when we are researching the past.

If you’d like to learn more about Madeline’s work, you can check out her website, which includes links to her 3D models and her book. You’ll find the link in the show notes along with the link to a full transcript of this episode

And if you’re curious to dig deeper and learn how Madeline got into archeology and the places it’s taken her, she shares that and more in a Patreon exclusive video.

Our Patreon is free to join and gives you access to all of our bonus content. If you’d like to support the podcast monetarily, we’d of course appreciate that too. Thanks so much for listening and cultivating your curiosity with this.

We’ll see you soon.

Credits

Madeline Robinson: Guest

Noor Hanania: Theme Music Creator

Laura Riveiro: Podcast Cover Art Creator

Brooke Norton: Production Support

Smiti Nathan: Host, Video Editor

Patreon Supporters

Thank you to our paid Patreon members for supporting the creation of this episode:

Nancy Fulton, Giancarlo Annese, Thomas Martin, Stuart A. Griffin, Tagg, Ann Brookens, Diana Madden, Cole Guinn, Seemingly Unrelated Podcast, Chris, GeometryGirl, John McGrath, Carmen Guerrero

Other Content You Might Enjoy