Principal Component Discriminant Analysis for Feature Extraction and Classification of Hyperspectral Images

feature extraction is one the most important subjects in the classification of hyperspectral images. It is necessary before classification and analysis of hyperspectral images. Principal component analysis (PCA) is one of the most conventional unsupervised feature extraction methods which extracts features with the largest power. PCA discards the components of data with small variance while components with small variance may have useful information for discrimination between classes in classification process. We propose to apply the linear discriminant analysis (LDA) to those components of PCA which have small power. So we extract the informative components for classification instead of discarding them. The proposed method that is called principal component discriminant analysis (PCDA) improves the classification accuracy and works better than both PCA and LDA. The experimental results obtained by using two hyperspectral data (an urban image and an agriculture image) are show the good efficiency of proposed method.