Hand-eye calibration for an unfocused light-field camera

A calibration framework is established for an unfocused light-field camera and a robotic arm. With Gaussian optics and light-field imaging principles, the mapping relationship between a point light source and its corresponding plenoptic disc feature is established, and the intrinsic and extrinsic parameters of the unfocused light-field camera are calculated through nonlinear optimization. Transformation matrices for eye-to-hand and eye-in-hand con-figurations are subsequently solved and are validated by applying them to an industrial light-field camera-robotic arm system. With the proposed calibration method, 3D reconstruction for calibration board in different poses is demonstrated and calibration uncertainty is discussed in detail. (c) 2022 Optica Publishing Group

Automated summary

A calibration framework is established for an unfocused light-field camera and a robotic arm, and the mapping relationship between a point light source and its corresponding plenoptic disc feature is established. The method calculates the intrinsic and extrinsic parameters of the camera and solves the transformation matrices for eye-to-hand and eye-in-hand configurations. The proposed calibration method enables 3D reconstruction of calibration boards in different poses and discusses calibration uncertainty in detail.