Dual-frequency pattern scheme for high-speed 3-D shape measurement

Area of Science:

• Optics and Photonics
• Computer Vision
• Metrology

Background:

• Phase unwrapping is crucial for 3D reconstruction but faces challenges like phase jumping and discontinuities.
• Conventional spatial and temporal phase unwrapping methods have limitations in accuracy and speed.
• Real-time 3D data processing requires efficient algorithms for phase generation and reconstruction.

Purpose of the Study:

• To develop a novel dual-frequency pattern for robust phase information and reduced phase unwrapping ambiguities.
• To overcome the shortcomings of conventional spatial phase unwrapping techniques.
• To achieve higher quality phase data with fewer patterns compared to temporal methods.
• To propose fast, look-up table based algorithms for real-time phase generation and 3D reconstruction.

Main Methods:

• A novel dual-frequency pattern combining high-frequency and unit-frequency sinusoid components was developed.
• The proposed pattern scheme addresses phase jumping and discontinuities in phase unwrapping.
• Look-up table based algorithms were developed for fast phase generation and 3D reconstruction.
• The algorithms were tested for real-time processing of a 640 x 480 video stream.

Main Results:

• The proposed pattern scheme successfully overcomes phase jumping and discontinuities.
• Higher quality phase data was achieved using fewer patterns compared to conventional temporal methods.
• Phase data generation reached 1063.8 frames per second.
• Full 3D coordinate point clouds were generated at 8.3 frames per second, significantly faster than previous studies.

Conclusions:

• The novel dual-frequency pattern scheme offers a superior approach to phase unwrapping for 3D reconstruction.
• The developed look-up table based algorithms enable high-speed, real-time 3D data processing.
• This advancement provides a significant speed improvement for phase measuring profilometry applications.