Fast terahertz reflection tomography using block-based compressed sensing

Area of Science:

• Physics
• Engineering
• Imaging Science

Background:

• Terahertz tomography is valuable for non-destructive imaging but is limited by long measurement times due to extensive data acquisition on 2D grids.
• Reducing measurement duration is critical for practical applications of terahertz tomography.

Purpose of the Study:

• To develop a novel fast terahertz reflection tomography technique.
• To reduce acquisition time and computational cost without sacrificing image quality.

Main Methods:

• Implementation of block-based compressed sensing directly in the spatial domain to decrease sampling points.
• Development of an overlap-average method to mitigate block artifacts inherent in block-based reconstruction.

Main Results:

• The proposed method significantly reduces measurement time in terahertz tomography.
• Block-based compressed sensing with the overlap-average method achieves substantial computational time reduction.
• Image quality is maintained despite the reduction in sampling points and computational load.

Conclusions:

• Block-based compressed sensing offers an effective strategy for accelerating terahertz reflection tomography.
• The developed technique successfully demonstrates fast and high-quality imaging of complex samples like integrated circuits and biological specimens.