Telephoto achromatic camera based on optical-digital co-design.

Area of Science:

• Optics
• Computational Imaging
• Deep Learning

Background:

• Correcting chromatic aberration (CA) in telephoto cameras is challenging, often requiring complex optical designs.
• Existing methods using simple optics and image algorithms struggle with compounded aberrations like defocus blur and multi-channel misalignment.
• Telephoto systems are prone to optical aberrations that degrade image quality.

Purpose of the Study:

• To develop an improved telephoto imaging system addressing limitations of current approaches.
• To mitigate axial and lateral chromatic aberrations (ACA, LCA) and other optical aberrations.
• To achieve high-quality imaging across the visible spectrum using a hybrid optical-computational method.

Main Methods:

• Leveraging distinct ACA and LCA characteristics across the visible spectrum in optical design.
• Implementing a cross-channel fitting method to suppress LCA.
• Utilizing deep learning for correcting ACA, defocus blur, and residual optical aberrations.

Main Results:

• The proposed approach effectively suppresses LCA using cross-channel fitting.
• Deep learning successfully corrects ACA, defocus blur, and other residual aberrations.
• Simulation experiments confirm the mitigation of CA and high-frequency information preservation in the green channel.

Conclusions:

• The developed telephoto imaging system effectively addresses compounded optical aberrations.
• The hybrid optical-deep learning approach delivers high-quality imaging results across the visible waveband.
• This method offers a viable solution for high-quality telephoto photography with simpler optical configurations.