Application of PCA and Hough Transform to classify features in optical images

Viewing fine features of object with optical instruments have become increasingly difficult as the dimensions of many features of interest have become smaller than the traditional optical resolution limit. Examples of these features can be found in semiconductor components and biological tissues. This has caused a move to non-optical methods such as scanning electron and atomic force microscopy techniques, or optical methods combined with signal processing techniques to provide clearer images of samples. This thesis presents a method to increase the resolution of an optical system. This is achieved by using principal component analysis (PCA). Once the PCA measured the object image parameters, the new clearer image can be reconstructed based on these parameters. This process works extremely well. Various aspects of samples measured by the PCA have been investigated, such as the shift of sample, the sample with different sizes, the orientation of sample and the impact of noise. These studies show that the technique is extremely robust, and has huge potential for general usage. The thesis also contains the detail of the Hough Transform technique which was used to provide the initial parameters to the PCA. From the analysis of the technique, it is concluded that the accurate measurement of the technique can be achieved by providing adequate templates of the object image for the system.EThOS - Electronic Theses Online ServiceGBUnited Kingdo