Single-pixel imaging using discrete Zernike moments.

Area of Science:

• Optics and Photonics
• Computational Imaging
• Image Processing

Background:

• Single-pixel imaging (SPI) offers a cost-effective alternative to traditional imaging systems.
• Zernike moments are valuable for analyzing image properties like rotation invariance.
• Compressive sampling enables efficient image acquisition with fewer measurements.

Purpose of the Study:

• To propose a novel single-pixel imaging technique utilizing discrete orthogonal Zernike moments.
• To demonstrate high-quality image reconstruction under compressive sampling conditions.
• To leverage Zernike moments for efficient and accurate object classification based on rotation invariance.

Main Methods:

• Illuminating the target object with two sets of Zernike basis patterns.
• Acquiring reflected light intensities using a single-pixel detector.
• Reconstructing the object image iteratively using Zernike polynomials and detected intensities.
• Classifying rotated images by measuring specific Zernike moment amplitudes.

Main Results:

• High-quality image reconstruction was achieved through theoretical and experimental validation.
• The proposed SPI system demonstrated effective object classification, grouping rotated images of the same content together.
• The classification method showed high efficiency and accuracy, attributed to SPI's speed and sensitivity.

Conclusions:

• The novel SPI technique effectively reconstructs images and classifies objects using Zernike moments.
• The rotation-invariant property of Zernike moments enables robust object classification.
• This method presents a promising approach for efficient and accurate imaging and classification tasks.