reference point/color for aero comparison

[Ciro Pabón]

This is an old software I saw several years ago, when it seemed cool, by Carnegie-Mellon, for "popups". Not in the league, but...

http://www.cs.uiuc.edu/homes/dhoiem/pro ... tware.html

A "little on thread", I posted eons ago something about pressure sensitive paint. Perhaps that's what the teams should use for "aero" comparisons...

VW new Beetle pressure coefficients at 50 m/s


Delta Wing, from 10 to 30 degrees (Quick Time required)

It is said that you can get 1.5 mbar precission and 0.5 seconds time lag.

If I weren't dying of a cold I would give a warm latino embrace to bazanaius, even if that approach to welcoming people gives some europeans the shivers... Anyway, welcome on board, man.

[bazanaius]

cheers for the welcome guys. I've been reading the site for quite a while but that was the first time I had anything to say!
vyselegend I think you're right. I havnt had a chance to collar my friend again but I think he meant that the video is simply a method of capturing a high number of 'regular' images (i.e. all same focal length, exposure time etc). I also think he implied by the way he spoke that the software has to be 'tuned' to whatever you're examining (be it faces, land masses or F1 cars).
If it is very feature specific I wonder how much it is worth the F1 teams investing in this and for what return? Whilst it would be very cool, I think that their current setups for the development of their own car would suit a different approach to modelling other people's cars. Simply get a journo to photo it and model it themselves in CAD or a windtunnel model.
In this way they can also tune it to their car - knowing the exact makeup of another car's part might not be that useful if the other car's overall aero profile is different.
B

[checkered]

A welcome from

me too, bazanaius, and thanks for Treead and Ciro for the interesting links. I've been meaning to post a response for quite some time, but am in a sort of a backlog situation with all the information to do with the subject. There's a lot of it ... I guess it'll take a little time to trickle through the synapses and get organised in my mind. At some point I'm sure I'll try some fairly low tech methods like back-engineering a wing from pictures with trial versions of labour intensive triangulation softwares.

Meanwhile, I have a second cousin who works as a 3D modeler and quite a while ago he showed me clips where he'd "immersed" virtual objects into handheld DV material ... this happens by pointing out fixed points in the video, whereafter the virtual objects' coordinates are automatically aligned according to the data. Not exactly the same thing, but it seemed easy enough, so there's another platform that could be expanded on. It must've been a 3DS plugin of some sort, seemed to work like a charm for him.

[Carlos]

Maybe we won't need 3D software. There is a new chip that combines the usual light/colour detection with optics, a chip with a layer of lens on top of an array of CDD pixels on a chip.
http://gizmodo.com/359199/new-camera-ch ... otos-in-3d

[vyselegend]

I don't really understand if it can really help modelling something in 3D, or just create a "3D effect"? Does it operate from a single viewpoint or could it "record" an object by revolving around it (in which case it should be able to aquire 3D data)

As I understand it, it's a bit like the stereographic system, but instead of a binocular type of lens it has something like 64 photosensors per cell. I don't get how many cells there are in the lens though...

OT- Did you read the comments? LOL the first thing that came up in their mind is weather it would allow 3D porn or not.

[Jersey Tom]

To really analyze a competitors car, can you imagine how DETAILED of a scan you'd have to get?

Digitizing 3d objects isn't childsplay, guys.. especially if you want to get some accurate and a file size thats manageable by any computer.. ie capturing actual curves rather than raw 3d digitized data.

It's a time consuming, challenging, and/or very hands-on operation.

[Carlos]

I think your right vyselegend, this chip would give a sort of stereographic image. It would need software to compare the multiple images from each individual lens to sort out all the overlapping bits and bytes and then output a compilation that would have an impression of depth. Sort of iimitating an insect eye which has multiple lens, which is then interpreted by a little bug brain. This chip won't give up a hologram, so it's just an improvement on a conventional camera.