magic4wei/kitti repository overview

⁠Usage

This image is built based on magic4wei/miniconda:miniconda3-latest-ubuntu<XX>.04-gui⁠.

⁠Download (Docker Pull)

Currently this repository contains images on Ubuntu 18.04 LTS and Ubuntu 20.04 LTS. Choose one to pull:

docker pull magic4wei/kitti:kitti2bag-py3.9-ubuntu18.04
docker pull magic4wei/kitti:kitti2bag-py3.9-ubuntu20.04

⁠Interactive Way

Run image (starting with base environment) and bind your dataset volume to the container:

docker run -it --rm \
--name kitti2bag \
-v <your_dataset_folder>:/opt/data \
magic4wei/kitti:kitti2bag-py3.9-ubuntu18.04

In the container, switch to KITTI environment by running

conda activate kitti

Then you can work with kitti2bag. In the docker run command above, dataset are mounted at /opt/data directory, so you can get into the directory and convert KITTI dataset to rosbag there.

⁠Noninteractive Way

You can also run kitti2bag commands directly when executing docker run command:

docker run -it --rm \
--name kitti2bag \
-v <your_dataset_folder>:/opt/data \
magic4wei/kitti:kitti2bag-py3.9-ubuntu18.04 \
kitti2bag -t 2011_09_26 -r 0002 raw_synced /opt/data

If you want to work on multiple dataset in one command, try to create a shell scripts in <your_dataset_folder> (which will be mounted at /opt/data in the container) and run that script. For example, here we create a script named work.sh in <your_dataset_folder>:

#!/bin/bash
SCRIPT_DIR=$( cd $( dirname $(readlink -f ${BASH_SOURCE[0]}) ) && pwd )
cd ${SCRIPT_DIR}
. ${MINICONDA_DIR}/etc/profile.d/conda.sh
conda activate ${ENV_NAME}
kitti2bag -t 2011_09_26 -r 0005 raw_synced .
# Add more tasks

Then we can run the commands below to execute work.sh at container start time which will convert KITTI dataset to rosbag we want in the same directory:

docker run -it --rm \
--name kitti2bag \
-v <your_dataset_folder>:/opt/data \
magic4wei/kitti:kitti2bag-py3.9-ubuntu18.04 \
/opt/data/work.sh

⁠kitti2bag Usage

See tomas789/kitti2bag⁠ for more details.