Skip to content

Latest commit

 

History

History
51 lines (31 loc) · 1.35 KB

visualizations.md

File metadata and controls

51 lines (31 loc) · 1.35 KB

Visualizations

Most visualizations from the paper are generated using the functions in the viz folder.

Point Cloud

Smaller values in the voxel-size result in fewer points. The VSCode extension Python PLY Preview is used to visualize the .ply file.

python3.8 viz/get_point_cloud.py --path /seed4d/data/static/Town01/ClearNoon/vehicle.mini.cooper_s/spawn_point_1/step_0 --voxel_size 0.6 --viewpoint nuscenes sphere

pointcloud

Sensor Overview

Currently only implemented to visualize views from the ego vehicle.

python3.8 viz/vis_all_sensors.py --path /seed4d/data/static/Town02/ClearNoon/vehicle.mini.cooper_s/spawn_point_1/step_0/ --category static

sensors_full sensors

Ego-Exo Views

Showing front left, front and front right first and then a randomly sampled exo views.

python3.8 viz/vis_ego_exo_views.py --path /seed4d/data/static --towns 01 --spawnpoints 4 8 15

sensors

Gif

python3.8 viz/vis_gifs.py --path /seed4d/data/dynamic/Town10HD/ClearNoon/vehicle.mini.cooper_s/spawn_point_4/step_0/ --vehicles 71

gif

Timesteps

python3.8 viz/vis_timesteps.py --path /seed4d/data/dynamic/Town10HD/ClearNoon/vehicle.mini.cooper_s/spawn_point_12

timesteps