Future of cave mapping is right around the corner?

[KeyserSoze]

Behold Project Tango. The idea of 3d cave mapping has been discussed here before and now it looks like Google has it working off of a smart phone. A the end of the video he mentions that they are currently distributing the devices to software developers. How can we get cavers involved in developing a 3d cave mapping program?

http://www.youtube.com/watch?v=Qe10ExwzCqk&feature=youtu.be

[Martin Sluka]

Behold Project Tango.

Anyway you'll need to measure the reference stations and in any way you'll need to add details manually. But fact, you will have ONLINE the exact shape of passage - plan and profile outline, may short the time of sketching in a cave.

[rlboyce]

Pretty freaking sweet. I wonder how much ambient lighting this device needs...

[caver.adam]

I imagine it will need about the same amount of light as a standard cell phone. The real questions is how well it will stitch together images that are dark for several seconds and then light later when you walk down the passage. If it can handle changes in lighting as the user walks, this will be the way to go for cave mapping! If you have a passage with walls that are too smooth for the sensor to differentiate you could use a couple of medium sized dodecahedrons on the floor for extra targets.

[Martin Sluka]

you could use a couple of medium sized dodecahedrons on the floor for extra targets.

I'm afraid the mirrors will not work. People makes 3D scans use fluorescent tape instead. The technologies of distance measurement used in the theodoliths and in the handy distancemeters is different. Second one are not able to measure distance od mirroring surfaces.

[Steven Johnson]

I imagine it will need about the same amount of light as a standard cell phone.

Yes and no: for purposes of cave mapping I suspect you really don't need distinct color channel information, just luminosity, which might give you leeway to work in a bit less light than otherwise. (That said: most cell phone cameras are especially terrible in low light, so, who knows.)

Here's a demo video of it being used in a (well-lit) room, and the resulting 3D detail:

http://www.theverge.com/2014/2/21/54339 ... ject-tango

Based on that, it certainly appears to be capturing enough data points to produce a point cloud that you could get a very serviceable map from.

There are plenty of questions, of course: what's the minimum light level needed for reasonable performance? How much passage can you record in local storage, and how much battery power will it take? Are there ways to mitigate inertial drift enough to get reasonable loop closure? etc.

[caver.adam]

Flourescent tape would be cheap and effective for edge detection in smooth passage. I like it.