Compact multimode fiber beam-shaping system based on GPU accelerated digital holography

Area of Science:

• Optics and Photonics
• Computational Imaging
• Fiber Optics

Background:

• Advanced beam shaping in multimode fibers is crucial for various optical applications.
• Current systems often rely on acousto-optic deflectors (AODs), adding complexity and cost.
• Graphics processing unit (GPU) technology offers significant computational power for real-time optical control.

Purpose of the Study:

• To develop a graphics processing unit (GPU)-based system for real-time beam shaping at the end of a multimode fiber.
• To eliminate the requirement for an acousto-optic deflector (AOD) in fiber-based beam shaping systems.
• To computationally remove interference effects for improved image quality.

Main Methods:

• Implementation of Gerchberg-Saxton and Yang-Gu algorithms on a GPU.
• Utilizing computational power for real-time image reconstruction and interference removal.
• Comparison of GPU implementation speed against traditional CPU implementations.

Main Results:

• Achieved real-time beam shaping without an acousto-optic deflector (AOD).
• Successfully removed undesired interference effects computationally.
• GPU implementation demonstrated a speed increase of two orders of magnitude compared to CPU.

Conclusions:

• A computationally efficient system for fiber-end beam shaping has been developed.
• The proposed method enables video-rate image control with minimal interference.
• This advancement simplifies complex optical setups and enhances performance.