Camera SDK Detection and Tracking

Post Reply
bnns_bc
Posts: 2
Joined: Tue Mar 27, 2018 7:03 am

Camera SDK Detection and Tracking

Post by bnns_bc »

I've been looking at the provided samples of the Camera SDK and I was wondering if it supports detection of patterns and tracking of patterns movement vertically, horizontally and distance from the camera?

I've looked in the 6DoF and the sample provided seems to track all objects/segments on screen and use that as estimated projection and it doesn't seem to track specific shape or pattern if anyone has suggestions or recommendations, documentation of how this can be achieved using the SDK, or any other tools, that I can look at, it would be greatly appreciated.
Chris.Johnson
Posts: 139
Joined: Tue Aug 15, 2017 1:56 pm

Re: Camera SDK Detection and Tracking

Post by Chris.Johnson »

Hello bnns_bc,

Please have a read through the wiki that goes over the provided developer tools provided by OptiTrack. This includes the Motive API, Camera SDK, and the NatNet SDK.

https://v20.wiki.optitrack.com/index.ph ... oper_Tools
All the Best,
Chris Johnson
Customer Support Engineer
NaturalPoint
bnns_bc
Posts: 2
Joined: Tue Mar 27, 2018 7:03 am

Re: Camera SDK Detection and Tracking

Post by bnns_bc »

I've read through the documentation as near as i could figure it:

Camera SDK: On it's own does not provide any interface for tracking but provides all sorts of functionality to controlling every aspect of the Camera. (It has helper functions for positioning 6DoF but no pattern detection just generalized tracking of markers)

NatNet SDK: Provides an interface for communication with realtime data over network protocols and by itself doesn't perform tracking.

Motive API: Is an all in one solution for multiple forms of tracking including rigid bodies.

The question i have relates to, is rigid body tracking and detection functionality, is it only available in the Motive API? Is there some way of tracking patterns in the Camera SDK by itself (without the Motive API)? How many rigidbodies can the Motive API from references i've seen it listed as 32 is that a software or harware limitation, is there any way to increase it to more than 32? How quick can detection be performed with a rigid body that start completely occluded (with the full amount of 32 rigid bodies in use with none of them currently visible)? What is the behaviour of a rigid body (of lets say 7 points) that starts occluded and is then has a portion of the markers revealed (5 points) where the revealed markers share a match with another rigid body (which is also not shown), will it pick the first one, try and map it both rigid bodies, or wait until a unique rigid body is found?

I thank you for your attention to this matter, any assistance you can provide will be greatly appreciated.
Post Reply