top of page

Underwater 3D modelling from wrecks to reefs

Written by Trey Cowher, commercial/technical diver and photogrammetry enthusiast.


Image of a scuba diver at the surface, before descending
Buddy check with students at the surface before descending for the final dive on a technical dive course. © Zach Benge

It is not often one stumbles upon a technology that has the ability to revolutionize the way we view, interact, and learn about our oceans. When you do though, it is not always obvious that the potential is there. The technology of photogrammetry is nothing new, but due to our rapid technological development over the past twenty years we can do much more with it now than could ever have been imagined.


Tech diver to photogrammetry student


close-up underwater image of a scuba diver with a rebreather device attached to their regulator
Diving a CCR closed-circuit rebreather allows for deeper depths and longer bottom times which leads to more data collection per dive. © Andrea Linhartová

My journey started in 2020, right before the COVID lockdown on the island of Koh Tao, Thailand. I was already a recreational dive instructor training to become a technical and project diver. Technical diving or tech diving is a form of diving that allows for longer, deeper, and planned decompression dives. Sometimes this means using multiple tanks with various breathing gasses or even utilizing closed circuit rebreathers which can greatly extend the bottom time of a dive. Project diving is a different scope of diving focused mainly on the completion of a specific task. Often this type of diving covers scientific, research, exploration, and data collection tasks. Using technical diving as a tool for project diving is a necessity, and showing the interest in one usually leads to the involvement of the other.


Luckily, I had started to rub elbows with people who were well versed in both which led to an invitation to join a project in the southern gulf of Thailand. The objective was to dive on the sunken submarine the U.S.S. Lagarto to capture thousands of photos. Those photos would then be used to recreate a 3D model of the sub for the Wisconsin Maritime Museum. The museum honors the Lagarto and the city of Manitowoc where the submarines were constructed.


I had no formal training in photography, photogrammetry, or archaeology but here I was able to use my skills to assist in a project that would bring marine heritage to thousands of people.

a scuba diver diving at a very deep depth, illuminated by their torch
Diving a CCR closed-circuit rebreather allows for deeper depths and longer bottom times which leads to more data collection per dive. © Andrea Linhartová

Sadly, the worldwide pandemic had other plans, and the expedition ended up being suspended. However, my interest was piqued and I began working with Mikko Paasi and together we began to learn the techniques behind capturing images for the photogrammetry process. 


Photogrammetry is the technique of taking colour photographs with digital cameras to create digital 3D models of a subject. Basically you take hundreds, if not thousands of photographs slightly overlapping one after the other changing angles as you go. The larger the area the more photos you require. Certain variables underwater like intricate subjects with tree-like complexity and poor visibility also increases the necessary amount of photos.


Once data collection is finished it is time for model recreation using sophisticated software. The goal is to produce a “scan” of the subject down to millimetre or sub-millimetre level precision. Essentially the process takes physical objects from the real world and creates a virtual 3D copy that you can view, measure, and share. Pretty cool, eh?


Digital model of a shipwreck
Scan of Suphairin 313, Koh Tao, Thailand scan. © Trey Cowher

After learning the intricacies of the software, it was time to start playing. So, I grabbed my GoPro and laptop and did some topside practice. Soon enough I understood the process sufficiently to get in the water and continue learning, taking pictures alongside others to create models. I was soon capturing every image and producing my own models. Small wrecks became bigger wrecks, and bigger wrecks became deeper wrecks. Quickly, I began to realize the ability this technology has to preserve underwater objects in time, both large and small.


A scuba diver with two tanks, recording a scan of a shipwreck
Completing a scan on the local dive site HTMS Suphairin 313 which was created as an artificial reef, Koh Tao, Thailand. © Andrea Linhartová

Not all of the learning process was a cakewalk though. The marine environment is an unforgiving one and provides many unique difficulties to the photogrammetry process. Full days were spent meticulously collecting data only to have taken the photos to the computer for processing and be disappointed. In order to get any successful result you need to make sure your dataset has full alignment. Remember how you need to overlap your pictures and record from different angles? Well sometimes photos are blurry and missing details which leads to misalignment in your model. If there is misalignment then your model will be incomplete or include gaps throughout. 


There were many repeated visits to dive sites to complete duplicate 2,000+ image scans. I had to make adjustments between dives, and often during the dive, to accommodate the ever changing environmental conditions.


The biggest challenges underwater are poor visibility, lack of light, change of light intensity, and an unsteady hand.

However, when I would come back to the computer after multiple attempts and finally create a tangible result, the feeling of accomplishment was second to none. After having taken roughly 200,000 photos for more than fifty unique scanning subjects, including entire dive sites, I can safely say that I still have lots to learn, but achieving a solid result is now more common than misaligned datasets.


Living wrecks vs living rocks


Scanning wrecks turned into scanning entire dive sites which led me to start exploring possibilities with corals. Entire dive sites became coral heads, coral heads became coral polyps. I am no scientist, but I am also no stranger to the scientific process and whilst modelling corals, an idea popped into my head regarding coral health monitoring.


What if we could implement monthly scans using photogrammetry to better track coral growth and decay?

Scientists could receive monthly data that is infinitely more useful than still two-dimensional photographs. Tracking the metrics of length, area, volume, colour, and diversity are all things that photogrammetry can provide. Visual data goes from being a two-dimensional picture of a small one square meter area to a real-world model of the sea floor easily ten thousand square meters in size.


A scuba diver scanning a wreck, carrying a torch and scanner
Completing a scan at Green Rock dive site, Koh Tao, Thailand. © Andrea Linhartová
Two 3D scans of piles of rocks
Completed scan of the Green Rock dive site, Koh Tao, Thailand. Left image shows a photorealistic version of what you would expect underwater and the right image shows a coloured model. © Trey Cowher

I wanted to put this idea into action and started pursuing options. Global Reef is a conservation organization in Koh Tao specializing in technological innovation for marine health, they were perfect to collaborate with to implement my idea. Piers Baillie, the founder of Global Reef, and I discussed the possibilities of using photogrammetry for coral health monitoring and the idea was a hit. So far, working alongside the stellar Global Reef, and Hydronauts Diving teams, and under the RAID dive agency, I have been able to get some baseline scans of coral heads, reef cubes, coral restoration sites, and even specimens of diseased and decaying coral. 


A close-up of corals
Colorful coral heads scatter the landscape at Sipidan Island, Malaysia. © Trey Cowher

We did not stop there though; we have worked to gather orthomosaic data which is part of the latter end of the photogrammetry process. These orthomosaics are two-dimensional panoramic pictures that encompass a large area to a high detail. Think of a large area like a map. These mosaics are used to determine the ecology of a coral reef. Instead of a dozen or so shots of different coral heads we can now swim one hundred meters in a straight line while scanning and produce a much better representation of the ecology of that reef than using previous methods.



a close up of an anemone and clown fish
Anemone fish and its anemone home on Sipidan Island, Malaysia. © Trey Cowher

Now we are getting somewhere in collecting useful data, but how can we maximize the benefits of these scans? The key question is: where should we be scanning? Monitoring natural reefs is essential because of their vast size, immense age, and vital role in supporting diverse marine species. However, natural reefs lack control, defined geometry, and a baseline for comprehensive study. Artificial reefs offer a solution to this challenge. 


Photogrammetry, used alongside coral research projects and artificial reefs, is an ideal match. We can now create baseline models of newly installed artificial reef structures and monitor coral growth or decay from day one of installation.

Generating a growth model every three months is as simple as diving to the site and conducting a scan. This technology can be applied to reef blocks, coral restoration structures, coral farms, and even wrecks, which transform into artificial reef structures over time. So, my research has brought me back to wrecks, not only to preserve their history this time, but also to observe their evolution into thriving reefs and document the positive ecological changes they bring.


close-up of a swimming sea turtle
This curious little sea turtle came up and posed for the camera mid-scan at Green Rock dive site, Koh Tao, Thailand. © Trey Cowher

The future with new tools, technology, and techniques


I have spent a lot of time thinking about other ways photogrammetry could benefit people and the planet alike. A major project I am currently working on is developing a smartphone application where users can access local dive site scans. Divers can plan dives more accurately, track marine life more precisely, and provide data to scientists more consistently than ever before. That is just the beginning. With more advanced software, cutting edge underwater tools, and a refined recording technique, there exists the possibility to gather data that can produce models with millimetre and submillimetre precision. Bringing in this information alongside local weather patterns, tides, currents, precipitation and other factors, we can now bring better data to the brilliant minds working on delivering solutions to our oceanic problems.


In this way, photogrammetry could revolutionize the way citizen science works; it could be as easy as uploading your data to an app.

So, where to get started? Koh Tao, Thailand, as a major diving capital of the world, seems a good place to start. I aim to release the app later this year, with the goal of aiding divers, scientists, and benefiting the planet.


3D model of large boulders on the seafloor
Completed scan of No Name Pinnacle dive site, Koh Tao, Thailand. © Trey Cowher

In writing this article, my aim is to demonstrate that photogrammetry offers numerous benefits and can be learned and implemented without formal training. Despite diving being notoriously expensive, starting your photogrammetry journey requires only basic gear—a mask, snorkel, fins, and an action camera. Living in the information age means learning is more accessible than ever before with most information just a YouTube video away.


With more people engaging in photogrammetry, from wrecks to reefs, we can collectively contribute to conservation efforts and make a significant impact.

This is just the tip of the iceberg for photogrammetry's potential. As we advance, we'll uncover even more ways to leverage this technology, expanding its diverse applications and, if used correctly, its scope for saving our planet. The future of photogrammetry holds limitless possibilities, but one thing is certain: we're on the right path.


IG: explorer_tre









Comments


bottom of page