Compressive sensing computational ghost imaging.

Area of Science:

• Computational imaging
• Optical physics
• Signal processing

Background:

• Ghost imaging reconstructs objects from intensity measurements.
• Spatial light modulators (SLMs) enable wave field coding.
• Compressive techniques reduce measurement requirements.

Purpose of the Study:

• To investigate computational ghost imaging using phase SLMs.
• To reconstruct transmission-mask amplitude objects.
• To develop efficient algorithms for limited measurements.

Main Methods:

• Utilizing a phase spatial light modulator (SLM) for wave field coding.
• Employing compressive sensing techniques for image reconstruction.
• Developing maximum likelihood algorithms for Poissonian and Gaussian noise models.
• Applying sparse and overcomplete object modeling.

Main Results:

• Successful image reconstruction with fewer measurements than image size.
• High accuracy and sharp imaging achieved through sparse modeling.
• An approximate Gaussian distribution model proved efficient for Poissonian observations.

Conclusions:

• Computational ghost imaging with phase SLMs and compressive techniques is effective.
• Sparse object representation is crucial for high-quality reconstruction.
• Gaussian approximation offers an efficient algorithmic approach for noisy data.