Image reconstruction using spectroscopic and hyperspectral information for compressive terahertz imaging

Area of Science:

• Physics
• Engineering
• Imaging Science

Background:

• Terahertz (THz) time-domain imaging is a rapidly developing field with significant potential.
• Conventional THz imaging systems face challenges with long scan times and complex designs.
• Compressed sensing (CS) offers a promising approach to reduce scan time and detector costs in THz imaging.

Purpose of the Study:

• To develop advanced reconstruction algorithms for compressed sensing THz imaging.
• To enhance the accuracy and quality of reconstructed images in THz time-domain imaging.
• To leverage prior information and multi-frequency data for improved THz image reconstruction.

Main Methods:

• Developed a single-band CS reconstruction method incorporating prior phase information and spatial correlations.
• Created a multi-frequency CS reconstruction method utilizing spatial sparsity and hyperspectral band correlations.
• Validated algorithms using experimental data from a single-pixel THz imaging system.

Main Results:

• The single-band method significantly improved reconstruction quality by using phase and amplitude-phase correlations.
• The multi-frequency method further enhanced image quality by exploiting spectral phase and cross-band correlations.
• Both algorithms demonstrated effective scene estimation from experimental THz data.

Conclusions:

• Advanced CS reconstruction algorithms can overcome limitations of traditional THz imaging.
• Utilizing spectral and spatial information is crucial for high-quality THz image reconstruction.
• The developed methods offer a pathway to more efficient and effective THz imaging systems.