Data reduction techniques of hyperspectral images: A comparative study

In hyperspectral image analysis, the Principal Component Analysis (PCA) is most commonly used dimensionality reduction (DR) technique. This feature extraction method is known as PCA-DR. But PCA suffers from high computational cost and large memory requirement while applying to hyperspectral image. In this paper, the dimensionality reduction is carried out using PCA and its variants such as Segmented-PCA (Seg-PCA), and Folded-PCA (F-PCA). These three methods have been compared in terms of peak signal to noise ratio (PSNR), structural similarity index (SSIM), execution time and memory requirement.