Underwater Photogrammetric Documentation as a Testbed for 3DF Zephyr Software: An Evaluation Study in the Search for an Alternative to Agisoft Metashape
Driven by the need to find alternatives to Agisoft Metashape due to geopolitical restrictions, this study evaluates 3DF Zephyr's capability in underwater photogrammetry and confirms its ability to produce high-quality 3D models comparable to industry standards, demonstrating that reconstruction quality depends primarily on input data rather than software limitations.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
The Digital Time Machine Under the Waves
Imagine you are an archaeologist, but instead of digging through dry dirt with a trowel, you are exploring a sunken world hidden beneath the surface of a lake. The problem? Water is a tricky neighbor. It swallows light, turning vibrant reds into ghostly blues and greys, and it's full of floating dust that makes everything look blurry. To study these underwater treasures—like old shipwrecks or lost tools—scientists use a technique called photogrammetry. Think of this as a digital magic trick: you take hundreds of overlapping photos from different angles, and a computer stitches them together to build a perfect, 3D virtual model of the object. It's like taking a flat puzzle and turning it into a solid statue you can spin around on your screen.
For years, the "gold standard" software for doing this underwater magic has been a program called Agisoft Metashape. It's the reliable workhorse that almost everyone uses because it has proven itself in countless difficult dives. But recently, things have gotten complicated. Because the company behind Metashape is based in Russia, and due to current global tensions and trade restrictions, many researchers in the West can no longer easily buy or use it. This has left the underwater archaeology community in a bit of a panic: "If we can't use our favorite tool, how do we keep documenting history?" They need a new, trustworthy software that can handle the messy, low-visibility reality of the deep without breaking a sweat.
The Great Software Swap: Testing the New Contender
This is exactly the question Wojciech Gajtkowski set out to answer in his research. He wanted to see if 3DF Zephyr, a software made by an Italian company, could step up and do the job that Metashape usually handles. Instead of just running a side-by-side test with the same photos (which is one way to compare tools), Gajtkowski decided to test 3DF Zephyr's "muscle" in the real world. He wanted to see if it could handle a messy, real-life dive on its own, regardless of what kind of camera the diver was holding.
To put the software to the test, Gajtkowski dove into Lake Kłodno in Poland, a place with murky water where visibility was only between 1.2 and 1.6 meters. His target was the wreck of a wooden fishing boat, about 4.93 meters long, resting at a depth of roughly 6.6 meters. He didn't just use one camera; he strapped two different types to a single frame to capture the scene simultaneously. One was a consumer-grade Insta360 Ace Pro action camera (the kind you might clip to a helmet), and the other was a professional Sony A7IV full-frame mirrorless camera.
He created four different "recipes" for the software to try:
- Model A: Standard 4K video from the action camera.
- Model B: Raw, flat-profile video from the action camera (giving the software more data to work with).
- Model C: High-resolution photos from the professional camera with a wide 16 mm lens.
- Model D: High-resolution photos from the professional camera with a 35 mm lens, plus extra color correction.
The results were like watching a chef cook with different ingredients. When 3DF Zephyr tried to build the model from the action camera footage (Models A and B), it did a decent job of showing the boat's shape, but the details were a bit fuzzy, and the whole thing had a strong green tint, just like looking through murky water. However, when the software used the photos from the professional camera (Models C and D), the results were stunning. The models were incredibly sharp, showing tiny cracks in the wood and clear details of the boat's structure. The 35 mm lens version (Model D) was the sharpest of all, though it looked a bit less "underwater" because the colors were so perfectly corrected.
The most important discovery here is that 3DF Zephyr didn't fail. It successfully built high-quality 3D models from both the cheap action camera and the expensive professional gear. The study suggests that the quality of the final 3D model depends much more on the quality of the photos you take (the camera and the lighting) than on the software itself. In other words, 3DF Zephyr is a capable tool that doesn't limit you; if you give it good data, it makes a great model.
However, the author is careful not to say this is a "perfect replacement" in every single way. He notes that he didn't run a strict, head-to-head race against Metashape using the exact same photos to measure tiny differences in geometric accuracy. He also points out some practical hurdles: the free version of 3DF Zephyr can't import existing 3D models, and the full license is quite expensive. Additionally, the files for these high-detail models are so huge that you can't easily upload them to popular sharing websites.
Ultimately, this study confirms that 3DF Zephyr is a fully functional, reliable tool for underwater archaeology. It proves that researchers who can't access Metashape due to restrictions have a viable alternative that can handle the difficult, low-visibility conditions of the deep. While more direct comparisons are needed to see if it matches Metashape's precision down to the millimeter, 3DF Zephyr has passed the first major test: it works, it's repeatable, and it can turn a blurry, green-tinted dive into a clear, digital treasure.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.