Comparison of Common Algorithms for Single-Pixel Imaging via Compressed Sensing.

Area of Science:

• Optics and photonics
• Signal processing
• Computational imaging

Background:

• Traditional imaging relies on detector arrays, facing limitations in resolution and cost.
• Single-pixel imaging (SPI) offers an alternative using a single-pixel detector.
• Compressed sensing (CS) enables efficient signal acquisition and reconstruction, applicable to SPI.

Purpose of the Study:

• To review the concept of compressed sensing in single-pixel imaging (CS-SPI).
• To summarize key measurement matrices and reconstruction algorithms for CS-SPI.
• To analyze the performance, advantages, and disadvantages of CS-SPI methods.

Main Methods:

• Illuminating targets with spatially resolved patterns.
• Compressively sampling reflected/transmitted intensities with a single-pixel detector.
• Reconstructing images using various CS algorithms and measurement matrices.

Main Results:

• Demonstrated the feasibility of CS-SPI through simulations and experiments.
• Evaluated the performance of different CS-SPI approaches.
• Summarized the trade-offs between various measurement matrices and reconstruction algorithms.

Conclusions:

• CS-SPI offers a powerful approach for overcoming traditional imaging limitations.
• The choice of measurement matrix and reconstruction algorithm significantly impacts CS-SPI performance.
• Future research directions for CS-SPI are identified.