Hi,
I checked through the site but couldn't find any information, please let me know if I have missed any threads.
I was just wondering if anyone is using or tried to integrate camera traps or BRUVs with any of the tech like structured light, lidar or laser triangulation to be able to accurately measure animals in the wild and potentially gather other information like growth or nutritional status.
The system in the new iPhone seems relatively easy but low accuracy, and 3d scanners are realively cheap now especially with the boom of 3d printers.
Adding a structured light illuminator should not be too complicated but for what I read is not great for moving objects.
Triangulation 3D scanners emitting a laser line are capable of scanning moving objects so could be more suitable...
MBARI just published a paper using laser to rebuild the 3d structure of underwater animals (giant larvacean) in a super cool but complex system on their ROV. https://www.mbari.org/deep-piv-3d-flow/
Any suggestions or idea?
Do you think is too complicated VS stereo photogrammetry?
4 July 2021 2:29pm
Just in case anyone else was wondering, BRUV is "Baited Remote Underwater Video". I had to google it and wade through a couple of Urban Dictionary links before I found the correct one:
Wikipedia
5 July 2021 2:47am
Hi Aponzo.
We've been working on the innards of camera traps for a while, but haven't seen anything like that. It may be possible to incorporate an LED projector with a camera trap for the structured light. If the animal needs rotation, that probably wouldn't be possible though. If you have a project in mind for something like this, let us know.
Akiba
16 July 2021 11:45am
I've seen this done with computer vision. I'm not sure of how accurate the results where but in the video it was showing species type and length of fish.
16 July 2021 2:35pm
Could you explain a bit more, do you want a measurement of length, or whole volume?
Like @wkcmark said, there are some simple solutions for measuring length fish, using stereo arrangements for example (2 cams looking at the same spot from slightly different angles). I also found this creative solution, using a single camera in a special acrylic case, that makes use of refraction indices. This, again, is for underwater systems.
If you could project two lasers at a known angle on the body of the animal like in exploration submarines, you could calculate the distance from the camera and measure the animal, although this would be a manual process.
Akiba
Freaklabs