Where am I?
Trying to scan underwater sites such as submerged landscapes or very broken wrecks often results in becoming hopelessly lost. Add into the mix the vision limiting experience of low visibility and a difficult job gets a lot harder. When on a sightseeing dive disorientation can be frustrating as you fin around missing the points of interest. But on a photogrammetry dive disorientation generally ends in frustration:
- Running over the same features more than once & shooting more images than you have to.
- Failing to scan key points of interest.
- Worse of all; gaps in the scan and batches of images that simply will not align.
Wouldn’t it be really cool to know where you had been, knew exactly where you were and going, and to mark objects of interest as you swim around? Even in zero or very limited visibility and having never dived the site before?
Secrets and Targets
Without scaling, ground control points (GCPs) and GPS the resulting models are ok, but models without measurement or scale don’t really deliver their potential. Using control points underwater is fairly straightforward but getting the GPS measurements to tie up needs more work. What is needed is a way to directly record GPS underwater but the radio signals never penetrate the water. For a while I have had a solution to this issue waiting for an opportunity…
I can’t say too much about the latest project but I can say it was in Greece and it will be on television in the near future. Researching sites like this offsite always reaps rewards and it became apparent this project was going cover a large area. The local diver who found the site would not be available on every day but did send an outline of the area. A quick analysis hinted at 16 acres of seabed. Even in clear water learning your way around would take many dives. With a tight time budget how could we locate areas of interest in such a large site? How could we reliably return to any features needing a second look?
In my previous work documenting the work of Royal Navy dive units I had photographed underwater navigation kit used to survey invasion beaches for mines. Something like that would do the job! The decision was taken to acquire some underwater GPS based navigation equipment with Northern Diver’s Nimrod – a handheld underwater navigation tool with desktop mission planning software – arriving the day before the project started.
With the mission planning software loaded and from the comfort of my office I started to plan the first dive. Waypoints were identified and the mission was to swim the perimeter of the site and establish its limits. That was going to be about 1km of swimming from the shore. The mission was loaded into Nimrod and everything packed away for the flight to Greece.
The first dive is always a bit of a shakedown. A chance to check kit and check everything works. This was also the first time I was using Nimrod underwater. The first target was 80m from the shore. Nimrod has several ways of displaying navigation but my favourite is the live map – yes a map – showing the waypoints relative to the current position. Swimming on a constant heading is easy and the distance started to count down. Close to the first waypoint I dropped a coded ground control point marker on the seabed and marked its location on Nimrod. As I was finning around it was apparent the natural feature navigation would be a struggle as (to the untrained eye) the seabed looked similar.
Back to the Start
After 85 minutes I had ticked off all the waypoints and was heading back when it occurred to me I was close to my coded control point. Could I find it again? Nimrod gave me a heading and a distance of 32m. As I approached the distance dropped …20m…10m…8m…6m…5m…3m…and I was starting to think I had lost it…then it appeared in the slight depression I had placed it. Just think about that for a minute: after 85 minutes of swimming around a site I have never seen before I swum a straight line to a randomly located control point.
I eventually surfaced after a fin across featureless sand and found myself facing the entry point. 1km later and I was back right where I started and had marked 6 targets of interest needing a second visit.
Without giving away too much of the forthcoming TV broadcast I think it fair to say using underwater GPS navigation for a diving & photogrammetry survey is a game changer.
With sites that have a wealth of sonar data such as Deeper Dorset’s P47-D crash site or the UB116 we simply merge the two sets of data and Agisoft Photoscan Pro works its magic. But for big sites without survey or prior knowledge where we have to learn the layout then Nimrod increases efficiency by a considerable factor. Everyone will have to wait to see it in action and find out more when the programme is aired but I know I will be working with Nimrod again.
Only one thing caused us to pause and question ourselves on the third day…I do hope the octopus that stole the coded targets enjoys their prize. But with accurate GPS navigation not even the marine life could stop us and the resulting model covers a large area and is the most accurate to date.