Multi-channel high-precision FPGA-embedded time correlated single photon counting systems for optical scattering...

Area of Science:

• Photonics and Optical Engineering
• Computer Vision
• Signal Processing

Background:

• Scattering imaging is crucial for industrial applications like remote sensing and autonomous driving.
• Existing time-correlated single-photon counting (TCSPC) systems face limitations in performance and linearity.
• Visualizing objects through scattering media remains a significant challenge.

Purpose of the Study:

• To develop a high-linearity, multi-channel TCSPC system with enhanced temporal resolution.
• To introduce a novel image reconstruction method for scattering imaging using deep learning.
• To demonstrate the system's capability in visualizing objects obscured by scattering media.

Main Methods:

• Implementation of a multi-channel TCSPC system on a Xilinx UltraScale FPGA.
• Achieved temporal resolution of 10 ps with high linearity.
• Developed a plug-and-play (PnP) image reconstruction framework integrating scattering models with FFDNet (2D) and FastDVDnet (3D) denoising neural networks.

Main Results:

• The proposed TCSPC system demonstrated a temporal resolution of 10 ps.
• Achieved differential nonlinearity within [-10, 9.9] ps and integral nonlinearity within [-14, 9.9] ps.
• Successfully visualized complex scenes obscured by scattering media through confocal imaging experiments and simulations.

Conclusions:

• The developed multi-channel TCSPC system offers superior performance for scattering imaging applications.
• The novel PnP reconstruction method significantly enhances image quality in scattering media.
• The integrated system shows great potential for industrial applications requiring advanced imaging capabilities.