Independent Component Analysis for Hyperspectral Imagery Plant Classification

In land investigation, it is often required to extract plant and vegetation information from land covers, especially when plants are sparsely dispersed. To avoid the expensive ground survey, hyperspectral remote sensing image is adopted due to its narrow spectral bandwidth and high spectral resolution. However conventional unsupervised classification techniques often suffer from requiring priori as input parameter and sensitiveness to interference. This paper proposes an Independent Component Analysis (ICA) based unsupervised classification algorithm. ICA is a technique that stems out from the Blind Source Separation. In hyperspectral data processing, ICA projects data vectors to the space where the items of the vectors are mutually statistically independent, and therefore is capable of extracting various kinds of plant information. So as to strengthen the contrast of the resulted independent components, histogram adjustment and mathematical morphology post-processing procedure are appended after ICA decomposition. Through real hyperspectral data experiments, our algorithm has been verified to have better performance for classification than K-means and ISODATA. Besides, computation efficiency and noise robustness have also been improved by a noise filtering preprocessing procedure