The study of harmful micro/nano biological organisms or other specimens has many potential applications in security and defense or novel disease detection. Therefore, developing reliable, automated, and low-cost methods for real-time sensing, visualization, and identification of harmful pathogens or biological cells are essential in combating catastrophic diseases or for new medical treatments. The optical imaging systems using digital holography have been extensively investigated for three-dimensional (3D) visualization and recognition of rigid objects. In contrast, biological specimens have dynamic activities such as moving, dividing and growing. Consequently, these issues make it difficult to three-dimensionally visualize biological specimens. Moreover, many specimens such as cell parts in protozoans, bacteria, and sperm tails are essentially fully transparent unless stained, killing the specimens. Accordingly, it is insufficient to use the 2D profile of them based on image intensity for their visualization or identification. Phase-contrast techniques have been developed for the non-invasive visualization of the transparent specimens because the difference in densities and composition within them give rise to changes in the phase of light passing through them.; This thesis presents a novel methodology for 3D sensing, imaging, and identification of micro/nano biological organisms. The proposed digital holographic systems provide a real-time 3D recording and visualization of specimens. Statistical pattern recognition algorithms are developed for the optimum classification of specimens. This methodology aims at identifying the 3D images reconstructed from the digital holograms of specimens and recognizing very minute differences in thickness, size, and shape. This methodology develops into the white light in-line (WLI) digital holographic system that has much simple, practical optical setup. The major advantages of the proposed WLI digital holographic system are that it can analyze specimen across the entire spectrum and requires very simple, practical optical setups for 3D recording. In addition, optimal spatial resolution can be obtained by using shorter or applicable wavelengths. The methodologies and algorithms addressed in this thesis are believed to be substantial progress for the automated identification of transparent biological specimens. To the best of the knowledge of the author, this is the first report on using 3D holographic imaging data for optoelectronic identification of micro/nano biological organisms.
展开▼
