3 ms·

They are not photos, that's made clear from the article, they are renderings. Think of it this way, the ROV goes down and uses a high resolution sonar instrume

They are not photos, that's made clear from the article, they are renderings. Think of it this way, the ROV goes down and uses a high resolution sonar instrument to see the 3D structure in front of it down to a centimeter/millimeter level of detail, at the same time it takes pictures. Then they take all that data and process it with a computer in order to generate basically a 3D textured model. Then they take renderings from that model and produce the images you've seen. The individual images from the ROVs would be much less interesting due to the very small field of view and the limitations of the lights. See also: https://en.wikipedia.org/wiki/Photogrammetry https://en.wikipedia.org/wiki/Photogrammetry

tgb 10y ago
In no way was it clear to me whether or not the images were the renderings that was produced, I'm aware of photogrammetry and the limitations of the underwater photographs (hence my question), I even have experience doing computer vision with underwater ROVs and their limited cameras (though ours had good field-of-view). The problem is that with their captions, I first assumed that it's a photo used for photogrammetry. Or rather, I thought it was a mosaic of photos, all of which were used for photogrammetry. If I were the NYT and didn't want these to be mistaken for photographs, I'd label them as "renderings of a photogrammetric model" or such. As it is, a "photogrammetric image" could be one of several things. In retrospect, there are some tell-tale signs (like the edge of the model where the ground is discontinuous).
- tgb 10y ago
  I can't edit anymore but I want to add that I regret my tone above, I think it's overly combative. I now think my original mistake was thinking that the model that the photogrammetry would generate would 'obviously' be a mesh or point-cloud without color/texture data and so the images being entirely computer rendered didn't fully cross my mind.
rebuilder 10y ago
To be pedantic, "photogrammetry" means the models would have been constructed from a dataset consisting of photos.

tgb 10y ago

In no way was it clear to me whether or not the images were the renderings that was produced, I'm aware of photogrammetry and the limitations of the underwater photographs (hence my question), I even have experience doing computer vision with underwater ROVs and their limited cameras (though ours had good field-of-view). The problem is that with their captions, I first assumed that it's a photo used for photogrammetry. Or rather, I thought it was a mosaic of photos, all of which were used for photogrammetry. If I were the NYT and didn't want these to be mistaken for photographs, I'd label them as "renderings of a photogrammetric model" or such. As it is, a "photogrammetric image" could be one of several things. In retrospect, there are some tell-tale signs (like the edge of the model where the ground is discontinuous).

tgb 10y ago
I can't edit anymore but I want to add that I regret my tone above, I think it's overly combative. I now think my original mistake was thinking that the model that the photogrammetry would generate would 'obviously' be a mesh or point-cloud without color/texture data and so the images being entirely computer rendered didn't fully cross my mind.

rebuilder 10y ago

To be pedantic, "photogrammetry" means the models would have been constructed from a dataset consisting of photos.