pcd-de-noising

PyTorch Lightning implementation of CNN-Based Lidar Point Cloud De-Noising in Adverse Weather. The original paper can be found on `arvix`_. The data used in the paper is available in the `PointCloudDeNoising repository`_.

Documentation and contributing guidelines can be found on `readthedocs`_.

Quick Start

Create a Conda enviroment with provided pcd-de-noising.yml file:

conda create -n pcd-de-noising --file pcd-de-noising.yml

Then activate the environment pcd-de-noising with:

conda activate pcd-de-noising

Unpack the data in the data directory:

tar -xvf data/5.zip
tar -xvf data/8.zip

Use tensorboard to monitor training progress:

tensorboard --logdir=log/Mistnet

Then you can run the train.ipynb notebook to quickly train, validate, and run inference. It is all setup with checkpoint loading and tensorboard logging.