Dimensionality Reduction of Hyperspectral Image Using Different Methods

摘要

The Hyperspectral images (HSI) are images obtained across the electromagnetic spectrum. Basically, images having a greater number of dimensions and complexity in processing and analyzing the data. As the number of dimensionalities increases, its accuracy gets decreases. Hence it is necessary to reduce the dimensionality by applying a preprocessing step. This HSI is widely used in industries and technology like remote sensing, seed viability study, biotechnology, environment monitoring, food, pharmaceuticals, medical diagnose, forensic, thin films, oil, and gas. There are different methods to reduce the dimensionality of these images like Principal component analysis (PCA), Weighted sparse graphbased (WSG), Curvilinear component analysis (CCA), Fractal based, Independent component analysis, Empirical mode and wavelets, Embedding, Band selection, Component analysis, Neighbourhood.