Pointcloud Labeler Dockerfile

This Dockerfile sets up an environment for using the Pointcloud Labeler and SuMa for pointcloud processing.

Features

• SuMa: A package for surfel-based mapping and 3D reconstruction.
• Point Labeler: A tool for manually labeling point clouds.

Setup

1. Clone this repository:

   git clone --recurse-submodules https://github.com/AIR-UFG/pointcloud_labeler.git
   cd pointcloud_labeler

2. Build the Docker image:

   docker build -t pointcloud_labeler Docker

3. Run the Docker container:

   To run the Docker container, utilize the provided run script with the following parameters:

   ./run.sh pointcloud_labeler --rm [--nvidia]

• <image-name>: The name you assigned to the Docker image during the build process.
• --rm: Automatically remove the container when it exits.
• --nvidia: Run the container with NVIDIA GPU support.

After running the container, a folder structure will be created within the repository directory:

pointcloud-files
├── bin-files
└── labels

This folder structure is linked to the /root/dataset/sequences/00/ directory within the container. The bin-files directory is used to store the .bin files of the point cloud data. The labels directory is used to store the labels files generated by the Point Labeler.

Usage

1. Prepare Point Cloud Data:

   Place your .bin files inside the bin-files directory in the repository. This directory is linked to the /root/dataset/sequences/00/velodyne directory within the container.

2. Run SuMa: The SuMa visualizer is used to calculate the poses of the point cloud data.

   • Within the container, navigate to SuMa's bin directory and run SuMa visualizer:

     cd /root/SuMa/bin
     ./visualizer

   • Open the first .bin file of the point cloud in the /root/dataset/sequences/00/velodyne folder.
   • Press the play button to calculate poses.
   • Save the poses.txt file to /root/dataset/sequences/00/.

   For more information on how to use SuMa, please refer to the SuMa documentation.

3. Run Point Labeler:

   • Within the container, navigate to the Point Labeler's bin directory and run the Point Labeler:

     cd /root/point_labeler/bin
     ./labeler

   • Open the /root/dataset/sequences/00/ folder inside the Point Labeler to start manually labeling the point cloud.

   For more information on how to use the Point Labeler, please refer to the Point Labeler documentation.

Customizations

• Aliases are added to facilitate common commands:
  • visualizer: Launches the SuMa visualizer.
  • labeler: Launches the Point Labeler tool.
• The run.bash script launches the Docker container with appropriate configurations for GUI display and volume mounts.

Credits

• This Dockerfile is based on instructions by Armin Niedermueller.
• Point Cloud Labeling Tool and and SuMa were created by Jens Behley.