Centroid of a cluster with respect to Kinect ?

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

Centroid of a cluster with respect to Kinect ?

faizan
Hii,

I am trying to compute the centroid of a cluster extracted from a planar surface with respect to kinect.
After computing the indices of the cluster, and extracting the clustered point cloud, I computer the centroid using Compute3DCentroid.

I am trying to estimate the position of the centroid with respect to kinect, I dont think compute3Dcentroid of the cluster will give me the clusters position wrt kinect. Am I wrong in saying this or I am missing something.

I am trying to understand the values which I am getting now.

Regards,
Faizan Shaik
Reply | Threaded
Open this post in threaded view
|

Re: Centroid of a cluster with respect to Kinect ?

Sergey
Administrator
Hi Faizan,

The return value of pcl::compute3DCentroid() is in the same coordinate frame as the supplied point cloud. Therefore if your clusters points are in Kinects frame (which is obviously the case), then the computed centroid is also in Kinects frame.

Regards,
Sergey
Reply | Threaded
Open this post in threaded view
|

Re: Centroid of a cluster with respect to Kinect ?

faizan
Thank you,
but whats the metric system followed here.
do they correspond to meters or cm.
Neither of them look correct to me.

Is there any distance transformation matrix that I need to multiple. Any clue on this.

~Faizan
Reply | Threaded
Open this post in threaded view
|

Re: Centroid of a cluster with respect to Kinect ?

Sergey
Administrator
Faizan,

All the distances are in meters, e.g. 2.3f is 230 centimeters.

The Kinect device should output correct distances/coordinates out-of-the-box. If you are using your own calibration parameters then this might be an issue.

Cheers,
Sergey