Search code examples
javagoogle-project-tangodepth-buffer

Google Tango: How is the point cloud confidence data can be used

I was working with the xyz data only till the updates of Tango API. I was wondering about the confidence data returned onPointCloudAvilable() callback, how can I benefit from this data?

are any point out of the range 0-1 is not a correct data? In another words, is it mean that if I eliminate points not in the range will give me more accurate point cloud depth data ?

Thank you

Solution

  • In practice:

    I have tried to check the values of C for each frame I get, but the values are always 1.0. Someone knows if this is really implemented or it will in future releases?

    In theory:

    According to the documentation:

    An array of packed {X, Y, Z, C} values.
{X, Y, Z} is a coordinate triplet (in meters).  
C is a confidence value, in the range of [0, 1], 
where 1 corresponds to full confidence.

    I haven't try yet but in theory all points are in the range [0,1]... being 1 the "best" points. So eliminating points lower than 0.5 (or 0.9) should reduce the noise of your acquisition.