Ca: Simple and Accurate LiDAR-RGB Calibration Using Small Common Markers

Area of Science:

• Robotics
• Computer Vision
• Sensor Fusion

Background:

• Accurate sensor fusion requires precise knowledge of relative sensor positions.
• Current LiDAR-RGB calibration methods often need complex targets, limiting usability in diverse environments.

Purpose of the Study:

• To develop a robust and simple LiDAR-RGB calibration method using common targets.
• To reduce human intervention in the calibration process.

Main Methods:

• Exploiting the flatness of a standard A3 chessboard target.
• Establishing correspondences between LiDAR and RGB sensor measurements.
• Utilizing nonlinear optimization for relative pose estimation.

Main Results:

• Achieved robust feature extraction from sparse LiDAR data.
• Demonstrated calibration performance on par with or superior to existing methods.
• Validated the approach through quantitative and comparative experiments.

Conclusions:

• The proposed method offers an effective and simplified solution for LiDAR-RGB calibration.
• This technique is suitable for scenarios where complex calibration targets are impractical.