TSDFVolume::integrate() merges new depth map into current TSDF space based on known view pose.
The camera params, depth factor and volume settings are the same at each frame.
It does this by comparing each voxel's depth to actual depth value of incoming frame.
After the integration each voxel contains:
- signed distance to surface and positive weight, if it's close to surface
- 1 and positive weight, if it was seen and known to be empty