HUGE balls (and a much larger capture area)

Post Reply
The_Nomad
Posts: 57
Joined: Tue Mar 04, 2008 9:55 am

HUGE balls (and a much larger capture area)

Post by The_Nomad »

Heya!

I've been having some discussions with a friend of mine who has been working on mocap systems professionally for a decade or so, and he came up with an interesting suggestion.

Is there any reason that it wouldn't be possible to significantly extend the capture volume of an Arena setup - albeit at the cost of some capture fidelity - by using MUCH larger markers, and then placing the cameras much further away?

I'm thinking 2" spheres here, and probably aiming for a 25' or so capture volume by placing the cameras about 40' apart. Obviously you'd need to solve the cable length issue, but other than that, is there an obvious reason it wouldn't work?
The_Nomad
Posts: 57
Joined: Tue Mar 04, 2008 9:55 am

Re: HUGE balls (and a much larger capture area)

Post by The_Nomad »

OK, I've now tried this out.

The system can certainly see the markers much further away - about double the range over which it can see the standard 5/8" markers.

We attached larger markers to the 3-point calibration wand, moved our system to an 8-camera configuration with the 4 tripods containing the cameras 25' apart, and attempted to wand.

Despite getting over 3,000 data points on each camera, the calibration wouldn't give us beyond a "poor" result.

We attempted to use that anyway, but the system subsequently couldn't find the ground plane.

Any suggestions or tips? Even "sorry, that's impossible" with an explanation would be good, to save us wasting more time on this.
NaturalPoint - Mike
Posts: 1896
Joined: Tue Feb 01, 2011 8:41 am
Location: Corvallis, OR

Re: HUGE balls (and a much larger capture area)

Post by NaturalPoint - Mike »

Cable length on V100R2 cameras is your limiting factor here. The maximum camera truss size for a usb system is about 22' square. With S250e cameras, however, we have been successful in creating much larger volumes due to significantly more flexibility in ethernet cameras combined with a higher resulution.
NaturalPoint - Mike
Posts: 1896
Joined: Tue Feb 01, 2011 8:41 am
Location: Corvallis, OR

Re: HUGE balls (and a much larger capture area)

Post by NaturalPoint - Mike »

Is 25' the square side or the diagonal?
The_Nomad
Posts: 57
Joined: Tue Mar 04, 2008 9:55 am

Re: HUGE balls (and a much larger capture area)

Post by The_Nomad »

25' was the square side.

We set the system up as follows:

2 cameras - 16' USB cables - OptiHub - 16' USB cables - 2 cameras

|

25' space

|

2 cameras - 16' USB cables - OptiHub - 16' USB cables - 2 cameras

However, as mentioned above, we had calibration problems - any suggestions there?
NaturalPoint - Mike
Posts: 1896
Joined: Tue Feb 01, 2011 8:41 am
Location: Corvallis, OR

Re: HUGE balls (and a much larger capture area)

Post by NaturalPoint - Mike »

Though not ideal, I did this about a week ago and it worked just fine. Can you attach a picture of your modified calibration wand with markers? (preferably with flash off). What's your cabling look like?
The_Nomad
Posts: 57
Joined: Tue Mar 04, 2008 9:55 am

Re: HUGE balls (and a much larger capture area)

Post by The_Nomad »

Ah, not very easily, I'm afraid - our mocap setup isn't terribly accessible without some notice! (Problems of being in the UK - there's a bit of a "large, accessible, cheap - pick two" issue around space here.)

We put velcro on the points where the markers usually attach, and attached the larger balls to them. AFAIK we were reasonably careful, but if that needs to be very exact it might have been a problem.

Cabling - it was, to be honest, probably under more tension than was ideal. We'd just rigged one camera at about 5', one at about 9', on each of the four tripods, with a hub in between two tripods on the ground. Is there anything specific about our cabling setup that would be a red flag for you?

Also, is there anything we could have been doing wrong in calibration?
Post Reply