A hand-eye calibration method using mutual viewing between two cameras

By employing a method of mutual observation between two cameras and iterative correction using Kalman filtering, the problems of high calibration cost and low efficiency in existing technologies are solved, achieving efficient online dual-camera calibration without the need for additional calibration materials, which is suitable for multi-camera systems.

CN117984314BActive Publication Date: 2026-06-30SOUTH CHINA AGRICULTURAL UNIVERSITY
View PDF 4 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
SOUTH CHINA AGRICULTURAL UNIVERSITY
Filing Date
2023-12-08
Publication Date
2026-06-30

AI Technical Summary

Technical Problem

Existing hand-eye calibration methods require precise calibration objects, are costly, are not suitable for harsh environments and multi-camera systems, have low calibration efficiency, and cannot be calibrated online.

Method used

A dual-camera mutual observation method is adopted. Through robotic arm movement and Kalman filter iterative correction, a loss function is constructed for online calibration. The camera itself is used as a calibration component, and point cloud matching is performed in combination with the ICP algorithm to achieve simultaneous calibration of the two cameras.

Benefits of technology

It can efficiently calibrate dual cameras online without the need for additional calibration materials, improving calibration accuracy and efficiency. It is suitable for multi-camera systems and reduces calibration costs and complexity.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117984314B_ABST
    Figure CN117984314B_ABST
Patent Text Reader

Abstract

This invention discloses a hand-eye calibration method using mutual observation between two cameras. The two cameras include an external camera (eye-on-hand camera) and an end-effector camera (eye-on-hand camera) mounted on a robotic arm, with the two cameras in a mutual observation posture. The external camera observes the end-effector, constructing the state transition equation and state observation equation for the end-effector's pose, and then using Kalman filtering to correct the end-effector's pose. The end-effector observes the external camera, utilizing the properties of coordinate transformation to construct the state transition equation and state observation equation for the external camera's pose, and then using Kalman filtering to correct the external camera's pose. A calibration loss function is then constructed to achieve hand-eye calibration using both cameras. This invention requires no additional calibration equipment, can simultaneously calibrate both cameras and perform online calibration, and achieves high calibration efficiency by iteratively correcting the calibration results using Kalman filtering.
Need to check novelty before this filing date? Find Prior Art