DIY and Entry-Level 3D-Scanner Forum

Hi All,

I am interested to see peoples thoughts on if it is possible or how complicated it would be to create a 3D scan of smoke in realtime (i.e. smoke would be moving).

I am interested in being able to measure the smokes density and thought that 3D scanning may be a good idea, as currently methods simply involve a single laser beam that measures how much the smoke obscures the laser beam, however this only provides information of average density of the smoke that is passing through the beam at the time, and is therefore not very detailed.

could 3D scanning give me a good 3d image of the smoke? or even a 2D slice of the smoke in a certain area of time? (i.e. 2d image changes over time as smoke passes through)

Any help/thoughts would be greatly appreciated!

Konrad

What you are talking about is making a 3D scan movie with each frame being a complete instant 3D capture. First, you would need smoke that was sufficiently dense to seem like a solid object. For the capture scans you would need a photogrammetry array of many, many SLR cameras capturing all sides of the smoke simultaneously, and shooting continuously to create the movie. Here is a video of Infinate Realities creation of a 3D scan face in motion. I'm sure they could do the job for you, but it would be incredibly expensive.

http://www.youtube.com/watch?v=H82wI9k9fDw&playnext=1&list=PL6BA3E8F9E4C7ABD1&feature=results_main

Hi Scantastic,

Thanks for the response,

However i am looking to be able to measure the density of the smoke (in a cross section maybe) and how it changes over time, I dont really need the pretty pictures, although that link you put up looks amazing! I just need the data that i can then analyse (this is for a research project).

I was thinking it may be able to be done using these flat beam lasers from different directions, so instead of just a single beam, it could measure the whole area, and using the lasers from different angles i may be able to 'triangulate' the position of the smoke..

But i am just guessing so please correct me if this would not be possible..

Konrad

Hi! I think you have two options: the first would be, as you mentioned above, to cut a section of the smoke cloud, for example with a bar laser and analyze the picture for the density of the smoke.
The other option would be with a kinect sensor. It sends out a infrared pattern. If this pattern is reflected by the smoke, you would see a 3d map of the cloud. However, I guess this method would just show the shape, but not the density of the cloud.

_________________
Bernhard
www.virtumake.com

To get internal details of the smoke you'll need to use rays which go through the smoke: Xrays for example.

Here's a link to Ben Krasnow's "DIY X-ray CT scanner controlled by an Arduino"
http://www.youtube.com/embed/hF3V-GHiJ78

Along the same lines but without the need for Xrays, you could use a translucent liquid in transparent one and compute the 3d structure from 2d images taken around the liquid chamber by using a similar algorithm to the one used above.