Capture volume for 8 camera setup

[trackback]

Hi,

I have a mobile platform that I want to track. There are 5 markers attached rigidly to the platform. I have a setup of 4 tripods with 2 Flex13 cameras on each of them (8 cameras totally) and I would like to know what capture volume is appropriate to get good tracking?

I understand the field of view of these cameras is 56 degrees so would a capture volume of about 12' x 12' be feasible? And if yes, then at what distance apart should I have the cameras (the tripods are set up in a square layout and I'm looking for the optimal value for the side of this square)?

Thank you.

[Shagan.Overman]

Hi,

Thankyou for choosing OptiTrack.

A 12'x12' capture volume is definitely feasible! I would recommend placing the cameras in the corners. Depending on how much occlusion from camera to marker visibility occurs the aiming will differ. If you wouldn't mind sharing an image of the mobile platform that would be helpful. Also any information on the amount of human traffic is expected in the space while the platform is in motion.

In the image below the cameras are aimed directly torwards the center of the space. The center blue square is 12x12. The red is 18x18. The lower cameras height can vary between 4'-6' if the higher one is 8.5' or taller. Be aware that if you walk too close to lower cameras tracking errors may occur.

HighOcclusion.jpg (78.38 KiB) Viewed 16198 times

If there is a minimal level of occlusion of markers occurring then the aiming of the cameras can be spread out to cover more space.

MinimalOcclusion.jpg (81.28 KiB) Viewed 16198 times

Let me know if you have any additional questions.

[trackback]

Thank you for your response and suggestions.

I currently set up the more conservative approach of cameras pointing to the center for my first data collect. Moreover, I have them pointing to the center at a certain height (and not the floor) since my markers are all at or near this height. I'm planning on trying the second approach you suggested later this week and compare.

I do not expect any occlusions from people or other objects in the area. The only occlusions will be from other parts of the platform itself or from one marker occluding the other. The platform is a 5' high object that moves around on wheels and we're looking to track its position and orientation.

The other question I had - I can actually place as many markers as I want on the platform and so are there some guidelines as to how many is optimal? I am going to group however many I place as 'one rigid body' in motive. Intuitively it seemed like the more markers, the better it is. But many markers also means that some of them won't be visible to all cameras at all times. And therefore I'm not sure what's the best.

Thank you.

[Shagan.Overman]

Happy to help!

If the structure you are attaching the markers to is rigid then you may place them anywhere the cameras will see them.

There is a default 4mm of tolerance to each expected marker position. If they deflect outside of that tolerance they will be considered occluded by the rigid body definition. All this to say... we don't want markers moving in relation to each other.

Five markers should be enough. You can feel free to add more as long as they aren't too close to one another. What we don't want is markers merging too often from a camera's perspective. The more markers that are added the more likely it is to happen. I wouldn't go above 6-7 markers in your use case. We fully expect occlusions to occur. They do not impact the accuracy of the rigid body.

Let me know if you have any additional questions!