Computational optical imaging: Applications in synthetic aperture imaging, phase retrieval, and digital holography

Dennis Joseph Lee, Purdue University

Abstract

Computational imaging has become an important field, as a merger of both algorithms and physical experiments. In the realm of microscopy and optical imaging, an important application is the problem of improving resolution, which is bounded by wavelength and numerical aperture according to the classic diffraction limit. We will investigate the resolution enhancement of phase objects such as transparent biological cells. One key challenge is how to measure phase experimentally. Standard interferometric techniques have the drawback of being sensitive to environmental vibrations and temperature fluctuations, and they use a reference arm which requires more space and cost. Non-holographic methods provide a way to overcome these disadvantages. Another challenge is how to reconstruct phase and amplitude from a digital hologram. The typical method of applying a filter in the frequency domain is limited by finite filter size. Optimization approaches offer a solution to this problem. The work presented here spans three main aspects of phase imaging microscopy including synthetic aperture imaging, phase retrieval, and digital holography. We develop a non-holographic microscope that uses off-axis illumination for resolution enhancement and demonstrate the first experimental measurements of referenceless phase retrieval at multiple angles. We implement a synthetic aperture microscope using an electrically tunable lens to defocus images, which avoids the need to mechanically move a camera on a translation stage. Finally, we improve the reconstruction of images from a digital hologram based on an iterative algorithm that alternatively updates amplitude and phase.

Degree

Ph.D.

Advisors

Weiner, Purdue University.

Subject Area

Engineering|Electrical engineering|Optics

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS