This repository contains code for parsing GoPro telemetry metadata to obtain GoPro images with synchronized IMU measurements. The GoPro visual-inertial data can then be saved in ros2 bag or EuRoC format. Thus, effectively paving the way for visual-inertial odometry/SLAM for GoPro cameras.
This repository use gpmf-parser from GoPro to extract metadata and timing information from GoPro cameras.
If you find the code useful in your research, please cite our paper:
@inproceedings{joshi_gopro_icra_2022,
author = {Bharat Joshi and Marios Xanthidis and Sharmin Rahman and Ioannis Rekleitis},
title = {High Definition, Inexpensive, Underwater Mapping},
booktitle = {IEEE International Conference on Robotics and Automation (ICRA)},
year = {2022},
pages = {1113-1121},
doi = {10.1109/ICRA46639.2022.9811695},
abbr = {ICRA},
bibtex_show = {true},
code = {https://github.com/AutonomousFieldRoboticsLab/gopro_ros},
}Tested on Ubuntu 24.04 (ROS2-Jazzy).
Building from source requires the FFmpeg development headers in addition to the
ffmpeg executable.
- First install ROS2 using this guide.
# First update package list
sudo apt-get update
# ROS 2 dependencies
sudo apt-get install -y \
ros-$ROS_DISTRO-ament-cmake \
ros-$ROS_DISTRO-rclcpp \
ros-$ROS_DISTRO-std-msgs \
ros-$ROS_DISTRO-geometry-msgs \
ros-$ROS_DISTRO-sensor-msgs \
ros-$ROS_DISTRO-cv-bridge \
ros-$ROS_DISTRO-rosbag2-cpp \
ros-$ROS_DISTRO-rosbag2-storage-default-plugins \
ros-$ROS_DISTRO-ros2launch
# System libraries and development headers
sudo apt-get install -y \
libeigen3-dev \
libopencv-dev \
libavcodec-dev \
libavdevice-dev \
libavfilter-dev \
libavformat-dev \
libavutil-dev \
libpostproc-dev \
libswresample-dev \
libswscale-dev \
ffmpegBefore proceeding, ensure all dependencies are installed. To install gopro_ros2 (currently save to bag file only):
mkdir -p ~/gopro_ros2_ws/src
cd gopro_ros2_ws/src
git clone https://github.com/AutonomousFieldRoboticsLab/gopro_ros2.git
cd ~/gopro_ros2_ws
colcon build --packages-select gopro_ros2 --symlink-install --cmake-args -DBUILD_GOPRO_TO_ASL=OFF
source ~/gopro_ros2_ws/install/setup.bash # Or add this to ~/.bashrc to make it permanentGoPro splits video into smaller chunks. By splitting up the video it reduces the chance of you losing all your footage if the file gets corrupted somehow. It’s called chaptering, and the idea is that if one chapter gets corrupted the others should still be okay because they’re separate files.
Both MCAP and SQLite3 storage backends are supported. The rosbag argument is
the output bag directory, not an output filename. Select the backend
explicitly with storage_id:=.mcap (the default) or storage_id:=.db3.
For example, write an MCAP bag with the default fast Zstandard compression:
ros2 launch gopro_ros2 gopro_to_rosbag.xml \
gopro_video:=/path/to/GX010001.MP4 \
rosbag:=/path/to/output/gopro_run \
storage_id:=.mcap \
mcap_compression:=zstd_fastThis creates the rosbag2 directory gopro_run/, containing metadata.yaml and
the generated .mcap storage file.
The supported MCAP compression profiles are zstd_fast (default),
zstd_small, and none. To write a SQLite3 bag instead:
ros2 launch gopro_ros2 gopro_to_rosbag.xml \
gopro_video:=/path/to/GX010001.MP4 \
rosbag:=/path/to/output/gopro_run \
storage_id:=.db3If you have multiple files from a single session, put all videos in same folder you can use the following command to concatenate into a single rosbag:
ros2 launch gopro_ros2 gopro_to_rosbag.xml \
gopro_folder:=/path/to/gopro_video_files \
multiple_files:=true \
rosbag:=/path/to/output/gopro_run \
storage_id:=.mcapTo save GoPro video with IMU measurements in Euroc format:
ros2 launch gopro_ros2 gopro_to_asl.xml gopro_video:=<gopro_video_file> asl_dir:=<asl_format_dir>If you have multiple files from a single session, put all videos in same folder you can use the following command extract all videos in a single folder:
ros2 launch gopro_ros2 gopro_to_asl.xml gopro_folder:=<folder_with_gopro_video_files> multiple_files:=true asl_dir:=<asl_format_dir>Enable save into EuRoC format.