Near-field terahertz single-pixel imaging with ultralow sampling ratio

Area of Science:

• Optics and Photonics
• Image Reconstruction
• Signal Processing

Background:

• Terahertz (THz) imaging is crucial for non-destructive analysis.
• Current methods face limitations in sampling ratios and acquisition speed.
• Efficient image reconstruction is key for practical THz applications.

Purpose of the Study:

• To introduce a novel reordered Hadamard basis method, Frequency-domain Modulus-correlation Hadamard (FMH).
• To demonstrate FMH's capability in reducing sampling ratios for high-quality image reconstruction.
• To enhance the acquisition rate of near-field terahertz single-pixel imaging.

Main Methods:

• Developed the Frequency-domain Modulus-correlation Hadamard (FMH) method with uniform spatial frequency increment.
• Integrated a pseudo-background post-processing technique.
• Compared FMH performance against Sylvester Hadamard and Walsh Hadamard methods using simulations and experiments.

Main Results:

• FMH significantly decreases the required sampling ratio for high-quality image reconstruction compared to existing Hadamard methods.
• Achieved a low sampling ratio of 4.27% for terahertz image reconstruction with a Pearson correlation coefficient > 0.9.
• Demonstrated a nearly 5-fold increase in acquisition rate for near-field terahertz single-pixel imaging.

Conclusions:

• The FMH method offers a substantial improvement in terahertz image reconstruction efficiency.
• FMH enables faster data acquisition in near-field terahertz single-pixel imaging.
• This technique has the potential to advance various THz imaging applications.