Image registration for LSPR study on gold nanoparticle.

• Main Author: Liang, Kaiying.
• Other Authors: Lee Kijoon
• Format: Final Year Project
• Language: English
• Published: 2012
• Subjects: DRNTU::Engineering::Nanotechnology
• Online Access: http://hdl.handle.net/10356/50160
• Institution: Nanyang Technological University

For the past decades, metallic nanoparticles have been the subject of intense research due to their unique size-dependent properties, which makes them superior and indispensable for a wide range of applications. In bio-related applications, the distinctive optical properties of metallic nanoparticles are employed to showcase its remarkable potential in labeling, delivery, heating and sensing. The derivation of these characteristic optical responses is attributed to the phenomenon of localized surface plasmon resonance, which varies with a nanoparticle’s composition, structure, size and shape. Hence, it would be a remarkable research breakthrough if the correlation between nanostructures and their optical responses can be established. In this report, the methodology of image registration is employed to identify this correlation. The data used in this project, are two images of a cluster of gold nanoparticles with varying structures obtained under an optical microscope and a scanning electron microscope. Two registration techniques namely Point Mapping without Feedback and Fourier Mellin Transformation are proposed and evaluated to identify the preferred technique for the registration of gold nanoparticles. The technique of Point Mapping without Feedback is highly reliant on the precision of the selected control points. Hence, it is unsuitable for registering nanoparticles which involves complex data and non-matching landmarks. With that, the author then proposed a technique on Fourier Mellin Transformation, whose resulted spectrum is invariant to rotation and translation. For a start, the author ventured into previous works to recover the possible scale factors relating the two images. Thereafter, the proposed technique was applied on the acquired scale factors, to determine the optimal scale factor. Lastly, the optical image’s area of interest was isolated and the SEM image was altered to maximize the areas of overlap.