Completed

Multimodal Sensor Calibration - Target based

Flexible extrinsic calibration using corner reflector with PnP/RANSAC/LM optimization

Multimodal Sensor Calibration - Target based

Overview

Advances in autonomous driving are inseparable from sensor fusion. Heterogeneous sensors are widely used for sensor fusion due to their complementary properties, with radar and camera being the most equipped sensors. To address the challenges of cumbersome operating procedures and well-designed experimental conditions, we propose a flexible, easy-to-reproduce and accurate method for extrinsic calibration of 3D radar and camera. The method places a single corner reflector (CR) on the ground to iteratively collect radar and camera data, obtain radar-camera point correspondences, solve the perspective-n-point (PnP) problem, and get the extrinsic calibration matrix using RANSAC and Levenberg-Marquardt optimization.

Key Highlights

  • Autonomous Driving
  • Sensor Fusion
  • Extrinsic Calibration
  • Flexible Method
  • High Accuracy

Methodology

Corner reflector target detection and PnP problem solving with RANSAC for outlier rejection and Levenberg-Marquardt optimization for parameter refinement.

Technologies

3D Radar Camera PnP Algorithm RANSAC Levenberg-Marquardt ROS

Project Details

Start Date
September 2022
End Date
April 2023
Status
Completed

Resources