Now that we have the depth output from the world, we need to use it a heuristic to construct a 3d world view. This should be relatively trivial since the depth is returned as a float from the matrix. Think about this as lidar reconstruction with monocular depth estimation. Look at some of Anjoo;s work with Nerfs, although i doubt we will need anything as complicated as that. I will be completely honest I don't think this has to be extremely accurate so feel free to use whatever transformation mechanism you want.
TODO:
[ ] make a camera.py file in the lib folder and write the starter code there
h3x4g0ns
commented
10 months ago
Now that we have the depth output from the world, we need to use it a heuristic to construct a 3d world view. This should be relatively trivial since the depth is returned as a float from the matrix. Think about this as lidar reconstruction with monocular depth estimation. Look at some of Anjoo;s work with Nerfs, although i doubt we will need anything as complicated as that. I will be completely honest I don't think this has to be extremely accurate so feel free to use whatever transformation mechanism you want.
TODO:
camera.pyfile in thelibfolder and write the starter code there