Using Bulldozer with Docker

Bulldozer is available as a Docker image, allowing users to run the full processing pipeline without installing Python dependencies locally.

You can use it either with a configuration file or with direct parameters.

This approach is particularly useful for: - Running Bulldozer without Python installation - Cloud or containerized workflows - HPC environments supporting containers (Docker / Apptainer)

Pull the Docker image

The official Bulldozer image is hosted on Docker Hub.

Download the image with:

docker pull cnes/bulldozer:latest

This image contains: - Bulldozer - All required Python dependencies - The full CLI environment

Usage with a configuration file

Bulldozer is executed inside the container while mounting a local directory containing the input data.

The directory is mounted inside the container as /data.

Linux / macOS

docker run --user $(id -u):$(id -g) -v <absolute/path>:/data cnes/bulldozer:latest /data/<config>.yaml

Windows (PowerShell)

docker run -v C:<absolute/path>:/data cnes/bulldozer:latest /data/<config>.yaml

Info

The --user option is required on Linux/macOS to avoid file permission issues. It is not needed on Windows due to the way Docker Desktop handles filesystem permissions.

Important notes

You must replace: - <absolute/path> with the absolute path to a directory on your system - <config>.yaml with your Bulldozer configuration file

Example directory structure:

project/
│
├── input_dsm.tif
├── config.yaml
└── output/

The container will see this directory as:

/data/

Configuration file requirement

The configuration file must reference paths inside the container.

Example configuration:

dsm_path: "/data/input_dsm.tif"
output_dir: "/data/output"

Warning

The output directory must also be located inside /data so that results are written back to the host system.

Example workflow

  1. Prepare a working directory:
my_project/
├── input_dsm.tif
├── config.yaml
└── results/
  1. Example configuration file:
dsm_path: "/data/input_dsm.tif"
output_dir: "/data/results"
generate_ndsm: true
nb_max_workers: 8
  1. Run Bulldozer:
docker run --user $(id -u):$(id -g) -v $(pwd)/my_project:/data cnes/bulldozer:latest /data/config.yaml

Results will appear in:

my_project/results/

Usage with direct parameters

Instead of using a configuration file, you can also pass CLI arguments directly.

Example:

docker run --user $(id -u):$(id -g) -v $(pwd):/data cnes/bulldozer:latest -dsm /data/input_dsm.tif -out /data/results -ndsm -workers 8

All CLI parameters available in the Bulldozer command-line interface are supported.

Output products

The results products are listed in the Bulldozer outputs page.

Performance considerations

When running Bulldozer in Docker: - Use -workers to control CPU usage. - Ensure sufficient RAM is available for large DSMs. - Use -mtd mem for faster processing on systems with sufficient RAM. - Mount input/output directories on fast storage.