Dimension reduction and its effects in hyperspectral data classification

In recent years dimension reduction and hyperspectral data imaging systems have gained significant attention from various research experts and institutions. By active acquisition of data through the use of data sensors, the middle infrared wavelengths can be altered and captured through various existing spectral channels from a particular defined interface within the surface of the earth. The hyperspectral sensors are quite crucial in the provision of detailed spectral information and this is quite priceless in as far as the increase in accuracy and the discrimination of materials of interest are concerned. In most instances, the hyperspectral solutions for imaging as well as the systems for spatial resolutions can be simply operated by application of sensors which ultimately enhance the process by playing two crucial roles; enhancing the provision of finer imaging details and enabling the analysis of even much smaller spatial structures that may be found within the images. As a consequence of consideration, it is invaluable to note the significant contribution of data classification and mapping in remote sensing and interpretation. One of the primary roles played by supervised classification of data is the use of classification input to observe or understand the implied spectral information about a given data. These includes information such as the value of the intensity for all the pixels used in achieving the grayscale images as well as the intensity vector for the RGB or even the images of high dimensions. This consequently leads to the production of data classification map which helps in the discrimination of different interests by the use of a representative sample. The central focus of this thesis, however, is to analyse the effect of dimension reduction in the said hyperspectral data classification systems. In the thesis, emphasis will be made on two important aspects of dimension reduction in hyperspectral data classification. These includes dimension reduction in PCA and ISOMAP then, decision making through the application of the Support Vector Machine (SVM), K-Nearest Neighbors (KNN) and the related MATLAB experiments for ISOMAP, PCA, SVM and KNN as desired. Keywords: Hyperspectral Imaging Data, Dimension Reduction, PCA, ISOMAP, SVM, KNN.