search

robodhruv / visualnav-transformer

Official code and checkpoint release for mobile robot foundation models: GNM, ViNT, and NoMaD.
http://general-navigation-models.github.io
MIT License
610 stars 78 forks source link

how to visualize the prediction traj and gt traj in image view with nomad train and eval pipeline? #33

Closed yancie-yjr closed 4 months ago

yancie-yjr commented 4 months ago

how to visualize the prediction traj and gt traj in image view with nomad train and eval pipeline?

kjctphr commented 4 months ago

+1

robodhruv commented 4 months ago

If you have the exact camera matrix for the datasets, you can use a simple transform for this visualization (example: https://github.com/robodhruv/visualnav-transformer/issues/22). The relevant code for this can be found in https://github.com/robodhruv/visualnav-transformer/blob/main/train/vint_train/visualizing/action_utils.py under the functions project_points() and plot_trajs_and_points_on_image().

Alternatively, you can produce a 2D top-down visualization.