Structured Light Three-Dimensional Measurement Based on Machine Learning.

Area of Science:

• Optics and Photonics
• Computer Vision
• Machine Learning

Background:

• Three-dimensional (3D) measurement using structured light is a vital technology across numerous industries.
• Traditional methods often involve complex phase-shifting steps, increasing operational complexity and computation time.

Purpose of the Study:

• To develop a machine learning-based approach for structured light 3D measurement.
• To eliminate the need for phase shifts, enabling real-time measurement capabilities.
• To improve the efficiency and accuracy of phase distribution recovery.

Main Methods:

• Utilized two distinct machine learning models to recover the phase distribution of a measured object.
• Implemented a grating-based structured light measurement system.
• Developed a novel method for one-dimensional phase wrapping, differing from traditional two-dimensional approaches.

Main Results:

• Achieved real-time measurement by eliminating the requirement for phase shifts.
• Reduced operational complexity and computation time significantly.
• Demonstrated a measurement error of less than 1%.

Conclusions:

• Machine learning effectively recovers phase distribution in structured light 3D measurement.
• The proposed method offers a simplified and faster alternative to conventional techniques.
• The system achieves high accuracy, making it suitable for practical industrial applications.