Theory, development, and application of quantitative phase imaging modalities on standard microscope platforms

Phase imaging;Optical transfer function;Transfer function inversion;Deconvolution;Fiber Bragg grating;Iterative optimization;Nonparaxial;Annular illumination

Bao, Yijun ; Gaylord, Thomas K. Electrical and Computer Engineering Cai, Wenshan Chang, Gee-Kung Valenta, Christopher R. Short, Martin Tibuleac, Sorin ; Gaylord, Thomas K.

University:Georgia Institute of Technology

Department:Electrical and Computer Engineering

关键词: Phase imaging;    Optical transfer function;    Transfer function inversion;    Deconvolution;    Fiber Bragg grating;    Iterative optimization;    Nonparaxial;    Annular illumination;   

Others  :  https://smartech.gatech.edu/bitstream/1853/61282/1/BAO-DISSERTATION-2019.pdf

美国|英语

来源: SMARTech Repository

【 摘 要 】

The object of this thesis is to develop and generalize quantitative phase imaging (QPI) methods to enable their more widespread use and their application to new classes of objects. Microscopic qualitative phase imaging has already produced impressive progress in biological and medical research. QPI is now being used even more widely in these existing fields as well as in industrial applications such as optical fiber characterization. QPI is not only quantitative in nature, but also label-free and thus able to image live cells in their natural, unperturbed environment. However, the conventional approach for QPI typically involves expensive custom stand-alone systems. To meet the growing QPI need and to reduce the cost, the Optics Laboratory has developed several new QPI methods that can be implemented on existing standard commercial microscope platforms. These methods include 1) 2D QPI method multifilter phase imaging with partially coherent light (MFPI-PC), 2) 2D QPI method phase optical transfer function recovery (POTFR), and 3) 3D QPI method tomographic deconvolution phase microscopy (TDPM). Since these methods have some limitations, the present thesis focuses on improving these methods.First, an analytical nonparaxial partially coherent 3D phase optical transfer function (POTF) was derived to describe the 3D image formation theory. Using this analytical nonparaxial 3D POTF, MFPI-PC was generalized to the nonparaxial condition without increasing computational time. In order to make MFPI-PC more suitable for annular illumination, weighted-least-squares MFPI-PC (WLS-MFPI-PC) was developed, in which a set of filters derived from least-squares fitting, further multiplied by an extra weight inversely proportional to the noise magnification factor, is used to replace the original binary filters. The analytical 3D POTF also greatly reduces the computation time needed in POTFR by making the transfer function semi-analytical. The improved MFPI-PC and POTFR have been compared through simulations. In addition, a unified, complete, and consistent description of the use of obliquity factor (OF) and OF modifications in 2D and 3D imaging of thin and thick objects was developed. In 3D QPI, an iterative regularization algorithm has been developed for TDPM, so that the refractive index can be reconstructed with high accuracy and with fewer rotation angles required, which enables faster measurements. An application of 3D QPI to fiber Bragg grating characterization was proposed by combining digital image processing techniques to overcome the short-period difficulty. Finally, specific future work is proposed, which includes further development of QPI methods as well as more applications.