forums.naturalpoint.com

In a research application, we are planning to take data from markers on a fairly large structure using 8 to 10 cameras. The area we need to capture is about 30 feet wide, 6 feet tall, and the field would be approximately 10 feet deep at the most. This is just an estimate, as we have yet to receive the exact dimensions of the structure and the room it will be in. I was wondering if this would be a feasible capture volume given 8 to 10 cameras and a good setup. I was also wondering whether active markers would be better than passive markers in such a large field. Thanks!

That capture area is currently beyond tha capabilities of the software/hardware.

30 foot area is just too large for the hardware to see across (would need a setup of at least 40-45ft.)

Ok, thanks for the quick response.

Vincent -- thanks for the help.

We were hoping to use IR-LEDs to enable the cameras to see the points from far away. Is the limitation the software, or the visibility of the markers?

Would there be any way to use enough cameras to "tile" the space, with several overlapping volumes?

(For instance, if you had a line of 32 cameras, all spaced one foot apart, could you have a capture volume 32' long, just not deep or tall?)

Thanks,
Jim

Jim,

Farther distances are not supported for two primary reasons. The first, as you mention, is reduced visibility of the markers due to reduced IR illumination. The other reason is the sensor resolution of the cameras is used to cover a much larger physical area when used with extended distances, this results in reduced resolution and accuracy (i.e. less pixels-per-inch). It may still be possible to track if you use active LED targets, but there are likely to be noticeable trade-offs.

While the Tracking Tools can operate multiple independent volumes at the same time, it does not currently support merging multiple volumes into a single contiguous one. If you can manage hand-off of tracked objects between the different volumes on your end then the Tracking Tools may be able to work for you.

So, would it be possible to have two separate volumes with 6 cameras, 3 cameras for each volume, and show the two separate volumes on one computer with Point Cloud tracking tools? Or is it necessary to have 2 computers, 3 cameras to a computer, and 2 versions of the software running to get 2 separate capture volumes? From what I understand, if we can calibrate the two volumes in the right way, we can simply translate the data we need in order to construct a visual model of both volumes combined.

Yes, you can run two separate volumes within one instance of the Tracking Tools Software. Using more than 3 cameras per individual volume might be a good idea.

Its possible to show transparent bounding boxes for each volume. Its also possible to offset one volume from the other volume so they are side by side for visualization and streaming.

The volumes can be physically overlapped if you want. If you overlap the volumes physically its important to consider that the only way you can align them is by setting their ground plane / orientation which isn't typically sufficient for good tracking.

For streaming, the data point cloud data from multiple volumes is combined and streamed out at 100fps. So, if the volumes are offset from each other then its easier to split the data back apart.

Here is an example screenshot from the Tracking Tools