RB2SV

rb2sv is a tool for extracting data from ROS bags and converting it to the Supervisely format.

Support

rb2sv now supports converting a ros2 bag to the following project types with message types listed below:

• Supervisely images project type
  • sensor_msgs/msg/Image
  • sensor_msgs/msg/CompressedImage
• Supervisely point_cloud_episodes project type
  • geometry_msgs/msg/PointCloud2

For images project type, the following message type can be converted to tags:

• sensor_msgs/msg/PoseStamped

However, for point_cloud_episodes project type, rb2sv does not support adding tags currently.

Requirements

Environments

1. Python ^3.10
2. ROS2 installed
3. Poetry installed

Install Python package dependencies with the following command:

poetry install

Usage

Clone the repo and install dependencies.

git clone https://github.com/NEWSLabNTU/rb2sv.git
cd rb2sv
poetry shell
poetry install

Source the ROS2 setup script before you start.

source /opt/ros/$ROS_DISTRO/setup.bash

Syntax

poetry run poe rb2sv [-h] [-q] -c TOOL_CONFIG.yaml
# e.g. poetry run poe rb2sv -c ./examples/example_config.yaml

• -c, --config-file-path: Required. The path to the configuration file.
• -q, --quiet: No logging during the conversion.
• -h, --help: Show this help message and exit

Format for the configuration file

The config file should be a yaml file with following keys:

• bag_path: Required. The path to the ros2 bag directory you want to convert.
• project_dir: Optional. The output directory.
• project_type: Required. Supports only images or point_cloud_episodes now.
• topic_pairs: Required. An array of pairs of (content-topic, tag-topic) to be converted.

See example_config.yaml for example configuration.

The output directory name is default to ./{rosbag-name}-supervisely. Each topic specified in the configuration yaml file would be treated as a dataset in the converted supervisely project.