Calibration issues

Post Reply
stepadou
Posts: 9
Joined: Wed Jun 16, 2010 12:10 pm

Calibration issues

Post by stepadou »

Dear all,

During my first encounter with an 8 camera Optitrack system I am already facing troubles with the calibration process. Wanding gives poor results and high lens distortion and then the 'Set Ground Plane' is not active, so I cannot proceed with it. Any ideas about troubleshooting?
Unfortunately, I haven't found any detailed manual about working on the Arena software...

Stella

P.S. I have also spotted an IR-LED that is not working on one of the cameras...
bsteagal
Posts: 322
Joined: Tue Jun 28, 2005 5:00 am
Location: Albuquerque, New Mexico

Re: Calibration issues

Post by bsteagal »

Hi Stella

Here's the link to Naturalpoint's tutorial videos, they are very complete and detailed on what to do:

http://www.naturalpoint.com/optitrack/p ... als/arena/

The things that will cause poor calibration results are poor positioning of the cameras (they are not all directed to the same area in your volume), make sure that you are keeping the wand visible to all cameras as much as possible when doing your calibration. then you have to perform the calculation prior to setting ground plane.

Make sure that you are not getting sunlight into your capture volume and that you have blocked any reflections that can cause problems.

As far as the IR-LED, Naturalpoint can have that camera replaced or repaired for you (hopefully you purchased it within the year)

But take a look at all the videos, they will be a lot more helpful than reading a manual

Hope this helps

Bryan
Bryan Steagall
Owner
Kidz Korner Studio
OptiTrack Distributor
Mexico, Central and South America
505-615-2410
bryan@kkstudio.us
www.kkstudio.us
Anakele
Posts: 28
Joined: Mon May 17, 2010 3:11 pm

Re: Calibration issues

Post by Anakele »

Hello Stella,
When you wand are you making sure to allow as many cameras to view the markers at the same time...in other words, are you keeping your body from blocking the cameras view of the markers while wanding? For that matter which wand are you using? Which software version are you using? If you are using the 3 marker approach have you switched to the 1 marker approach? How are your cameras oriented (2 in each corner)? How large is your capture volume? When you wand are you doing it in circular or round criss-crossing motions? For the camera that appears to have a malfunctioning IR-LED...try to calibrate and just focus on that one camera. See if you can make the entire screen green. If you can then there is nothing wrong with the camera. Note, you are not doing this to calibrate. You are only doing this to make sure that the camera is working properly. When you try to set the ground plane have you made sure that no other markers are visible in any of the cameras. The ground plane will not be set until ONLY the calibration square is visible in the capture volume.

That's about all I can think of. I wish you luck.
Best Regards,
Andre O'Brien
Anakele Vision and Design
stepadou
Posts: 9
Joined: Wed Jun 16, 2010 12:10 pm

Re: Calibration issues

Post by stepadou »

Hello to both!
Thanks a lot for the quick responses.

So:
(1) I was having problems with the calibration, because my cameras are wide (3.5) and I hadn't adjusted the settings.
(2) The camera is showing some inconsistency in its function (sometimes only 1 led will be on!), so I hope they replace it - the system was bought 1 month ago!
(3) I am using a 3-marker wand and the latest Arena software - is that wrong?
(4) I am still getting some deformation of the cameras' grids after calibration, but mainly on the edges - is this acceptible?

The setup of the 8 cameras is pretty weird...
I am trying to track mainly hand and finger movements of people seated on the floor (not changing position).
I have a 6-camera circular rig, 4 of which are at approx. 1 meter height, 2 (the front ones) are at approx. 1.7m height and I have also 2 additional cameras in the front which are very close to the floor, to pick the hand movements happening very low (rest of hands on the kness etc). The two low, front ones may be causing this deformation during wanding. (Perhaps I should send a snapshot of the setup).

What I need to understand is now:
(1) How I can only track particular points on the body (UPPER-body). My impression is that Arena only allows a particular sceletton setup. Is it something that I can only do in MotionBuilder (we have a license for it).
(2) Why I am getting so much noise when I add the markers on the person. While he is moving, all these different points shine (and disappear) at wrong places. I don't have any reflections or other light sources. All is very controlled in the lab.

I am using: 3 markers on wrists, 2 markers on elbows, 3 markers on shoulders, 3 markers on the back, the head-band and 3 little (face-) markers on each nail of the fingers.

I thank you very much again!

Best,
Stella
VincentG
Posts: 7728
Joined: Mon Jul 17, 2006 5:00 am
Location: Corvallis, Oregon

Re: Calibration issues

Post by VincentG »

OK, I think you have a couple of things going on.

1 - With that small of a capture area, cameras at approx 5ft and lower, you could very well be picking up the IR light bouncing off the actor itself.

2 - You can do a partial skeleton, within Arena, but it must include the hip markers.

3 - As far as the IR led goes, are you refering to the IR LED ring, or the numeric LED display?

4 - When you are calibrating, make sure that you get as much of the camera view covered with calibration data as possible.
stepadou
Posts: 9
Joined: Wed Jun 16, 2010 12:10 pm

Re: Calibration issues

Post by stepadou »

Hi there. It is so nice to get so immediate responses. I am really pleased about this!

1. I have set the intensity and other settings, so that the actor doesn't reflect at the distance he is sitting. This noise appears only when I have many markers on him; not during calibration or wanding for example...! Any ideas?

2. Thanks!

3. I am talking about the IR LED ring (the IR light emitters), not the display.

4. Yes, I am indeed. Can it be due to the lower cameras possibly seeing only a part of the space due to their positioning and distance?
Post Reply