Study of computational sensing using frequency-domain compression.

Area of Science:

• Optics and Photonics
• Computational Imaging
• Signal Processing

Background:

• Extending computational sensing from spatial to temporal domains enables capturing fast light signals with slow photodetectors.
• Traditional temporal computational sensing methods often require numerous measurements for acceptable signal-to-noise ratio (SNR) reconstruction.
• Slow detectors limit temporal resolution, posing challenges for capturing high-speed optical phenomena.

Purpose of the Study:

• To investigate a frequency-domain acquisition technique for capturing nanosecond temporal objects with a low-bandwidth detector.
• To enhance the signal-to-noise ratio (SNR) in temporal computational sensing.
• To enable sub-Nyquist sampling for efficient data acquisition and reconstruction.

Main Methods:

• Utilized a frequency-domain acquisition technique for temporal signal capture.
• Leveraged Fourier spectrum analysis for SNR gain (N, where N is the number of Fourier spectrum points).
• Employed sub-Nyquist sampling enabled by data compressibility and Fourier basis properties.

Main Results:

• Achieved a significant SNR gain (N) using the frequency-domain acquisition technique.
• Demonstrated the feasibility of reconstructing nanosecond temporal objects with a ten Hertz detection bandwidth.
• Showcased robustness against temporal distortions inherent in experimental setups.

Conclusions:

• The frequency-domain acquisition technique offers a robust solution for temporal computational sensing with slow detectors.
• This approach significantly improves SNR and allows for efficient data acquisition through sub-Nyquist sampling.
• The method provides immunity to temporal distortions, making it valuable for high-speed optical signal processing.