I have taken a recording of a blank wall to test the depth consistency of the L515 camera. The recording was collected using the capture function of the MATLAB wrapper that has been adapted to store the data.
I have taken an average of all of the depth points in the field of view after removing any anomalous points. I would assume that given a gaussian distribution of the errors in the depth measurements that averaging the points would give a consistent depth over time. However, I observe an obvious pattern in the depth data over 900 Frames. Is there a known cause for this i.e. fluctations in the laser intensity or temperature of the camera? I am trying to find a away to remove this low frequency signal but it is roughly the same frequency as the movement I am try to collect so filtering is not an option.
Please sign in to leave a comment.