Spatial-temporal binarization encoding basis modulation Fourier single-pixel imaging

Area of Science:

• Optics and Photonics
• Computational Imaging

Background:

• Fourier single-pixel imaging (FSI) faces limitations in switching rates and binary encoding of sinusoidal features.
• Existing FSI methods struggle with maintaining high reconstruction quality and efficiency.

Purpose of the Study:

• To present a novel FSI method using spatial-temporal binary encoding to overcome current limitations.
• To improve image reconstruction quality, speed, and detail preservation in FSI.

Main Methods:

• Developed an innovative image extension algorithm for scene preprocessing.
• Decomposed grayscale Fourier basis patterns into binary patterns using dynamic thresholding and error diffusion.
• Implemented a weighted reconstruction strategy for efficient single-pixel measurements.

Main Results:

• Achieved substantial improvements in image reconstruction quality (e.g., 80% SSIM increase in simple scenes).
• Demonstrated significant performance gains in complex scenes (e.g., up to 27.9% SSIM improvement).
• Reduced sampling time by up to 28.225s while maintaining comparable reconstruction quality.

Conclusions:

• The proposed spatial-temporal binary encoding FSI method offers superior performance over existing techniques.
• The method excels in preserving texture details and enhancing background reconstruction.
• This approach presents a versatile and practical advancement for diverse FSI applications.