Tyler and I first dove the PBM Mariner a couple weeks ago. During that dive, we had very limited visibility so taking pictures was challenging and I built a photogrammetry model based on those photos. It came out relatively well but was incomplete and I didn’t get all of the pictures needed to “link up” the different sections of the wreck.
This week after we “rescued” Lora’s anchor and 200+ feet of chain from the Cape Charles (that was quite the adventure), Tyler and I went back to the PBM hoping for better visibility. We got what we asked for!
Below are two photos from a similar angle between the two different days. Both photos are completely unchanged from what came out of the camera:
You might notice in the above pictures that the one on the left has more color and the one on the right doesn’t. Unfortunately, I had BOTH of my lights fail on the second dive on the PBM and so I had to take all pictures with natural light.
More on that in a bit…
Natural Light Photos
Before I get to the photogrammetry model, it is worthwhile to post some of the individual photos from the dive. The benefit of natural light photos is that backscatter is usually minimized. Below are some photos of the engine and prop on the port side that are a highlight the wreck (remember that the wing is upside down but facing “forward” so the “port” side of the wing is on the “right” looking forward).
The other exciting development is that Tyler spent some time searching around and found the remnants of a tire which helps confirm the identify of the wreck.
Below are some pictures of the other engine and prop:
Below are some screenshots of the model I built from the first dive. The “benefit” of these are that they were lit by artificial light so they have some good color. The downside is that I wasn’t able to tie everything together so they are “pieces” of the entire scene but don’t show the relative size or position.
My ultimate goal was to combine the photos from the first and the second dive to build a complete model.
However, I think that the difference in colors between artificial light and natural light photos doesn’t allow the Metashape software to correlate photos. I tried to align two photos that are almost exactly the same position but the software wouldn’t properly align them. I can only conclude that it is due to color differences.
I even tried to make matches after converting the photos of the tail sections to black-and-white and it still wouldn’t work. More research is required.
In the meantime, here is a model of the overall wreck site with all photos being in natural light (hence the “green” color shade).
Below is a screen capture of the model indicating the different areas:
Below are some screenshots of the model.
There are two follow-up objectives:
- Build a complete photogrammetry model with proper lighting. I’m not sure yet how or when I’m going to do this but it is “on the list.”
- More exiting, after we completed the second dive and Tyler had found the tire and some other debris, we scanned some of the surrounding area with the sonar on the Marissa and found what could be another large section of the wreck. We need to return and investigate….