Volume Calculation Module

[benjaminwinger]

No description provided.

[MurdochLabs]

Space carving looks like an interesting theorem... link here,
http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.163.250 , http://stackoverflow.com/questions/7161209/determine-the-volume-of-an-object-from-pictures is a stack overflow question on the topic

[MurdochLabs]

Look at calculating volume by point cloud: changes in pixels can be used to calculate changes in height, then perform shitty integration to determine more accurate volume.

[MurdochLabs]

http://stackoverflow.com/questions/20314556/generating-point-cloud-from-many-2d-images
Point cloud question; top comment has some good papers and stuff.

[MurdochLabs]

Stuff we're discovered so far: RANSAC is solvable by opencv, with some proper set up, so I'm inclined towards that.
WE can also use a program called Visual SFM http://ccwu.me/vsfm/ that I think can create 3d reconstructions and be used to test our stuff, at least as a baseline. Need to sit down and diagram this out super hard as well, though... Do that when we have time soon.

[MurdochLabs]

Was just looking at CONOPS, doesn't look like this is technically still necessary

[benjaminwinger]

True, though I think it's worth keeping around and working on as a long-term part of the project since they'll probably re-introduce volume calculation in the future (though it makes me wonder if the only reason they had it last time was because of the company sponsoring the software).