Before you can use the renderer in pyrender/

Use colmap to extract points from a sequence of images.

See how to do so here. Notably, you must save the sparse and dense reconstructions as text in separate folders for the next step.

Generate a scene from a COLMAP

After converting a sequence to 3D points using COLMAP, saving both the sparse and dense point clouds as text, run convertColmap.py from ./utils and follow the prompts. Note that before you do you will need to have the sparse and dense point cloud from the previous step.

Generate a workspace with that scene.

Described in pyrender/README.md.