lidar_pcl_features

LiDAR point cloud segmentation with classic ML approaches.

Getting started

1. Follow instructions to install Poetry:

   # Unix/MacOs installation
   curl -sSL https://install.python-poetry.org | python3 -

2. Check that poetry was installed successfully:

   poetry --version

3. Setup workspace:

   make setup_ws

4. Setup ClearML:

   clearml-init

5. Migrate dataset to your ClearML workspace:

   make migrate_dataset

6. (Optional) Configure and run Jupyter lab:

   make jupyterlab_start

Train

make run_training

Results

🧪 Experiment link

              precision    recall  f1-score
           0       0.52      0.73      0.61
           1       0.99      0.85      0.91
           2       0.35      0.50      0.41
           3       0.89      0.94      0.91
           4       0.45      0.25      0.32
    accuracy                           0.82
   macro avg       0.64      0.65      0.63
weighted avg       0.83      0.82      0.82