ROS Bag Exact Dataset

This repository contains scripts for extracting image and GPS data from ROS bag files. The primary script, rosbag_exact_dataset.py, is a combination of separate extraction scripts to simplify the process of working with image topics and GPS data in ROS bags.

Purpose

The script rosbag_exact_dataset.py:

• Extracts images from specified topics within a ROS bag file.
• Extracts GPS data from specified topics.
• Interpolates the GPS data to match image timestamps.
• Saves the extracted data in a well-organized output directory structure.

This is particularly useful for generating datasets from ROS bags for machine learning, computer vision, or mapping purposes.

Usage

To run rosbag_exact_dataset.py, ensure that you are in a ROS environment or a Docker container with ROS installed. The script is designed to work with ROS bag files and will fail without the required ROS dependencies.

Example command: python rosbag_exact_dataset.py

Known Issues

NumPy and Pandas Compatibility

There is a known issue related to the deprecation of np.bool in recent versions of NumPy. If you encounter the following error:

AttributeError: module 'numpy' has no attribute 'bool'.
np.bool was a deprecated alias for the builtin bool.

You can resolve this issue by upgrading numpy and pandas to compatible versions:

pip install --upgrade numpy pandas

For more details, refer to the official NumPy release notes: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations.

Files

• rosbag_exact_dataset.py: The main script combining image and GPS extraction along with GPS interpolation.
• rosbag_convert_gps_to_kml.py: A script to convert GPS data from ROS bag files into KML format.
• rosbag_exact_gps.py: A script to extract only GPS data from ROS bag files.
• rosbag_exact_image.py: A script to extract only image data from ROS bag files.
• rosbag_interpolate_gps_for_image.py: A script to interpolate GPS data to match image timestamps.

Here's an improved version of the markdown formatted for direct use on GitHub:

Run within ROS Docker

This section outlines the steps to run the necessary scripts within a Docker container with ROS installed.

Prepare the Container

1. Stop and remove any existing containers to avoid conflicts:

   docker stop ros_container
   docker rm ros_container

2. Start the ROS container with the required volumes:

   docker run -it --name ros_container \
   -v /robodata/ARL_SARA/GQ-dataset/bagfiles:/rosbags \
   -v /home/zhiyundeng/rosbag-tools:/rosbag-tools \
   -v /robodata/ARL_SARA/GQ-dataset/extracted_data:/extracted_data \
   --user $(id -u):$(id -g) \
   osrf/ros:noetic-desktop-full

Set Up ROS Environment

3. In one terminal, create a temporary ROS home directory and start the ROS master:

   mkdir -p /tmp/ros_home
   export ROS_HOME=/tmp/ros_home
   nohup roscore > /tmp/roscore_nohup.log 2>&1 &

4. In another terminal, access the container:

   # List all running containers (optional)
   docker ps
   
   # Open a bash shell inside the container as root
   docker exec -it --user root ros_container bash

5. Source the ROS environment:

   source /opt/ros/noetic/setup.bash

Working with ROS Bags

6. Check the information of a ROS bag file:

   rosbag info /rosbags/2024-08-12-11-32-46.bag

7. Save the rosbag info into a text file:

   ./save_rosbag_info_into_txt.sh

8. Extract dataset information from ROS bags:

   nohup python3 rosbag_extract_dataset.py &

Monitoring

9. Check the status of the running Python script:

   ps aux | grep rosbag_extract_dataset.py

10. Display the output of the running script:

    tail -f /rosbag-tools/nohup.out
    tail -f /extracted_data/extraction.log

License

This project is open-source and available under the MIT License.