By using the MRTK_V2 in Unity 2018.4.12f, I try to spatially map a multi-floor building with a first generation hololens. The hololens provides spatial mapping data that I wish to render at a remote location (a computer running a unity application outside of the scanned building) in near real-time (about 10 seconds). The idea of doing this is that someone at the remote location can follow the scanning process and see the 3D model of the building 'build up' in front of his/her eyes, while the hololens user can also be tracked in the 3D model.
I already have the mapping and tracking functionality, and i am able to send serialized meshes through a socket and deserialize them on the other end. However, I'm unable to render the meshes in the right position at the receiving side.
I tried transforming the vertice points of the individual meshes from local space to world space before sending the mesh, but this doesn't work as the position of the mesh is controlled by the world anchor that is connected to it. Disconnecting the anchor seems not favourable, as this would cause a lot of drift in the data. I would like to connect all meshes to a single point in space, to convert them to world space, but in my current understanding of the spatial mapping concept that isn't possible.
Now i'm wondering if I can render the meshes relative to the world anchors at the receiving side, for example by making use of the azure spatial anchors. I can imagine that this is not possible, as the anchors might need a connection to the real world. Also, the anchors do change their positions at runtime, therefore requiring constant updates, making the use of anchors less than ideal.
Any help would be greatly appreciated!
