Two-wavelength holographic micro-endoscopy.

Area of Science:

• Optical Engineering
• Metrology
• Biomedical Imaging

Background:

• Accurate micro-scale surface topography measurement is crucial for various scientific and industrial applications.
• Existing micro-endoscopic techniques face limitations in resolution and accuracy due to factors like fiber optic properties and noise.

Purpose of the Study:

• To present a novel micro-endoscopic topography measurement method using two-wavelength holography.
• To assess the impact of fiber optic characteristics and introduce virtual surface roughness as a resolution limiter.
• To develop and evaluate a compact micro-endoscope system with enhanced measurement accuracy.

Main Methods:

• Evaluation of inter-core dispersion and cross-talk in commercial imaging fiber bundles (CFBs).
• Design and fabrication of a micro-endoscope (<450 µm diameter) using 3D-printed micro-optics.
• Implementation of two-wavelength holography for topography measurement.
• Application of a simple phase unwrapping approach and a deep learning method for error reduction.

Main Results:

• Characterization of CFB properties and the concept of virtual surface roughness.
• Achieved a relative measurement error of 7.5% with a basic phase unwrapping method on a pyramid test object.
• Demonstrated significant error reduction using a deep learning approach tailored for noisy phase maps.
• The standard deviation of measurement error was reduced by 4.2 times with the deep learning method.

Conclusions:

• The developed micro-endoscope system enables accurate topography measurements at the micro-scale.
• Two-wavelength holography combined with advanced processing (deep learning) offers a powerful solution for high-resolution metrology.
• The deep learning approach effectively mitigates noise and improves the reliability of micro-endoscopic measurements.