Three-Dimensional Imaging Based on Refractive Camera Model and Error Calibration for Risley-Prism Imaging System.

Area of Science:

• Optics and Photonics
• Computer Vision
• Robotics

Background:

• Existing 3D reconstruction technologies face limitations in accuracy, detection range, and system size.
• Applications in intelligent driving and underwater exploration highlight the need for improved 3D imaging solutions.

Purpose of the Study:

• To develop a compact 3D imaging system and reconstruction method addressing current technological deficiencies.
• To enhance the accuracy and reliability of three-dimensional detection and reconstruction.

Main Methods:

• Utilized a compact Risley-prism 3D imaging system for multi-viewpoint imaging via prism rotation.
• Established a refractive camera model integrating pinhole and Snell's law for precise beam trajectory.
• Developed a forward projection method based on Fermat's principle and incorporated error calibration via optimization iteration.

Main Results:

• Analyzed systematic error influences through detailed simulations.
• Demonstrated significant enhancement in 3D reconstruction accuracy.
• Reduced the Normalized Root Mean Square Error (NRMSE) from 0.9076 to 0.0207.

Conclusions:

• The proposed compact Risley-prism system and refractive camera model offer superior 3D reconstruction capabilities.
• The error calibration method effectively mitigates systematic errors, improving overall reconstruction quality.
• This approach provides a robust solution for applications demanding high-accuracy 3D imaging.