ROS2 Hand-Eye Calibration

A ROS2 package for performing hand-eye calibration using a chessboard pattern. The package processes ROS2 bags, with topics from an RGB camera and robot poses, to compute the transformation between the camera and robot end-effector.

Prerequisites

• Docker
• ROS2 bag containing:
  • RGB images (/rgb/image_raw)
  • Camera info (/rgb/camera_info)
  • TF transforms between base_link and robot end-effector

Quick Start

1. Create a folder for your calibration bags:

   mkdir bag_folder

2. Copy your ROS2 bag files into the folder:

   cp /path/to/your/rosbag/* bag_folder/

3. Run the calibration. You have two options:

   Simple usage (with default parameters):

   cd docker
   ./run_docker.sh <your_bag_name>

   Full usage (customize all parameters):

   cd docker
   ./run_docker.sh <your_bag_name> \
       --image_topic /rgb/image_raw \
       --camera_info_topic /rgb/camera_info \
       --source_frame base_link \
       --target_frame right_arm_tool0 \
       --output_dir hand_eye_calibration_data \
       --collection_rate 20 \
       --chessboard_rows 4 \
       --chessboard_cols 7 \
       --square_size 0.065

4. Find results in the output directory (default: hand_eye_calibration_data):

   • hand_eye_calibration_results.json: Contains the calibration matrix and transformation

Parameters

All parameters are optional and have default values:

• image_topic — (default: /rgb/image_raw) RGB camera image topic
• camera_info_topic — (default: /rgb/camera_info) RGB camera info topic
• source_frame — (default: base_link) Robot base frame
• target_frame — (default: right_arm_tool0) Robot end-effector frame
• output_dir — (default: hand_eye_calibration_data) Directory for calibration results
• collection_rate — (default: 10) Collect every Nth frame
• chessboard_rows — (default: 4) Number of internal corners per row
• chessboard_cols — (default: 7) Number of internal corners per column
• square_size — (default: 0.065) Chessboard square size in meters

Output

The calibration results include:

• Rotation matrix
• Translation vector
• Quaternion representation