Stereo Camera Head-Eye Calibration Based on Minimum Variance Approach Using Surface Normal Vectors.

Area of Science:

• Robotics
• Computer Vision
• Calibration Techniques

Background:

• Head-eye calibration is crucial for robot vision systems, enabling accurate interaction with the environment.
• Existing methods can be complex and computationally intensive, limiting their application in real-time systems.
• Humanoid robots require precise visual-kinematic integration for tasks like manipulation and navigation.

Purpose of the Study:

• To develop an intuitive and simple head-eye calibration method for stereo cameras.
• To achieve globally optimal transformation estimation between a robot's head and eye.
• To enhance the efficiency and accuracy of calibration for humanoid robot vision systems.

Main Methods:

• An extended minimum variance approach is proposed for head-eye calibration.
• Surface normal vectors are utilized instead of traditional 3D point sets for calibration.
• The method incorporates both positional and orientational error variances in the calibration process.

Main Results:

• The proposed method demonstrates high accuracy in determining the head-eye transformation.
• Experimental results with synthetic and real data validate the method's efficiency.
• The technique offers a simplified approach compared to existing complex calibration procedures.

Conclusions:

• The developed stereo camera-based head-eye calibration method is accurate and efficient.
• Its intuitive nature makes it suitable for vision systems in humanoid robots.
• The use of surface normal vectors provides a robust alternative for calibration.