GNSS VINS ROS Bag Creator (Python version)

A Python tool to convert timestamp-named image files and IMU data into a ROS bag file. This utility simplifies the process of creating ROS bags from raw sensor data for robotics and autonomous systems development.

Overview

This tool allows you to:

• Convert a folder of timestamp-named images into ROS Image messages
• Convert IMU data from CSV files into ROS IMU messages
• Convert GNSS data from CSV files into ROS NavSatFix messages
• Package all sensor data into a single, time-synchronized ROS bag file

The resulting ROS bag can be used for simulation, algorithm testing, and sensor fusion development.

Requirements

• ROS (tested with ROS Noetic, but should work with other versions)
• Python 3.6 or higher
• OpenCV
• NumPy
• Pandas
• cv_bridge
• rosbag

Installation

1. Clone this repository:

git clone https://github.com/RuijieXu0408/gnss_vins_rosbag_creator.git
cd gnss_vins_rosbag_creator

1. Make the script executable:

chmod +x bag_creator.py

1. Install required dependencies:

pip install numpy pandas opencv-python 

1. Ensure your ROS environment is set up:

source /opt/ros/noetic/setup.bash  # Replace "noetic" with your ROS distribution

Data Format Requirements

Directory Structure

Recommended data structure:

dataset/
├── cam0/             # Folder containing timestamp-named images
│   ├── 1705396656652587231.png
│   ├── 1705396656752587432.png
│   └── ...
├── imu/
│   └── imu0.csv      # IMU sensor data
└── gnss/
    └── gnss0.csv     # GNSS/GPS data

Image Files

• Images should be named with nanosecond-precision Unix timestamps, e.g., 1705396656652587231.png
• Supported formats: PNG, JPG, JPEG
• All images should be in the same folder

IMU Data (CSV)

The IMU CSV file should contain the following columns:

• First column: timestamp (nanoseconds)
• ax, ay, az: Linear acceleration (m/s²)
• gx, gy, gz: Angular velocity (rad/s)
• Optional: qw, qx, qy, qz: Orientation quaternion

Example:

timestamp,ax,ay,az,gx,gy,gz
1705396656652587231,0.1,0.2,0.3,0.01,0.02,0.03
1705396656752587432,0.11,0.22,0.31,0.011,0.021,0.031

GNSS Data (CSV)

The GNSS CSV file should contain the following columns:

• First column: timestamp (nanoseconds)
• latitude, longitude, altitude: Position data
• Optional: status, service: GNSS status and service information

Example:

timestamp,latitude,longitude,altitude
1705396656652587231,40.7128,-74.0060,10.5
1705396656752587432,40.7129,-74.0061,10.6

Usage

Basic usage:

./bag_creator.py --output dataset.bag --image_folder dataset/cam0 --imu_csv dataset/imu/imu0.csv

Full options:

./bag_creator.py \
  --output dataset.bag \
  --image_folder dataset/cam0 \
  --image_topic /cam0/image_raw \
  --imu_csv dataset/imu/imu0.csv \
  --imu_topic /imu0 \
  --gnss_csv dataset/gnss/gnss0.csv \
  --gnss_topic /ublox_driver/receiver_lla

Command-line Arguments

Argument	Description	Default
--output	Output ROS bag file path	(Required)
--image_folder	Folder containing timestamp-named images	/cam0/image_raw
--image_topic	ROS topic name for image messages	/cam0/image_raw
--imu_csv	Path to IMU data CSV file	imu0.csv
--imu_topic	ROS topic name for IMU messages	/imu0
--gnss_csv	Path to GNSS data CSV file	(Optional)
--gnss_topic	ROS topic name for GNSS messages	/ublox_driver/receiver_lla

Verifying the Created ROS Bag

You can verify the contents of your newly created ROS bag using the following commands:

# Display bag information
rosbag info dataset.bag

# Play back the bag
rosbag play dataset.bag

# In another terminal, list available topics
rostopic list

# View messages on a specific topic
rostopic echo /cam0/image_raw
rostopic echo /imu0
rostopic echo /ublox_driver/receiver_lla

Tips for Large Datasets

• Processing large datasets can be memory-intensive. Consider splitting the conversion into smaller batches if needed.
• For image-heavy datasets, you may want to downsample the images or process only a subset to reduce file size.
• If your original timestamps are in a different format, modify the parse_timestamp function accordingly.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• Thanks to the ROS community for providing the tools and libraries that make this possible.