Co-phase errors simultaneous detection for optical sparse aperture systems via deep learning.

Area of Science:

• Optical Engineering
• Image Processing
• Machine Learning

Background:

• Co-phasing is critical for high resolution in optical sparse aperture systems.
• Existing methods struggle with simultaneous detection of piston and tip-tilt errors, often requiring step-by-step analysis and reducing efficiency.
• Tip-tilt errors complicate piston error detection in practical scenarios.

Purpose of the Study:

• To develop a novel method for simultaneous detection of piston and tip-tilt co-phase errors.
• To improve the efficiency and accuracy of co-phasing in optical sparse aperture systems.
• To overcome the limitations of existing single-type error detection methods.

Main Methods:

• Theoretical derivation of the relationship between optical transfer function (OTF) and co-phase errors.
• Utilizing an object-independent feature map (FM) derived from OTF.
• Employing a deep learning-based separation network to isolate piston error from tip-tilt interference.
• Integrating original and separated FMs into a detection network for simultaneous error identification.

Main Results:

• Demonstrated that tip-tilt error detection is independent of piston error.
• Successfully separated piston error from tip-tilt interference using a dedicated network.
• Achieved simultaneous detection of both piston and tip-tilt errors with high accuracy.
• The trained network requires only a single original FM input for detection.

Conclusions:

• The proposed deep learning method enables efficient and accurate simultaneous detection of piston and tip-tilt errors.
• This novel approach enhances the co-phasing process in optical sparse aperture systems.
• The method shows robust performance and high detection accuracy in simulations.