Feature Extraction Combining PCA and Immune Clonal Selection for Hyperspectral Remote Sensing Image Classification

A new feature extraction method based on immune clonal selection (ICSA) and PCA is proposed for classification of hyperspectral remote sensing image. As the hyperspectral remote sensing image is acquired in very narrow spectral channels, the resulting high-dimensional feature sets may contain redundant information. Therefore, feature extraction is necessary to classify a data with large dimension such as the hyperspectral remote sensing image. PCA is popularly used for feature extraction. However, in traditional PCA, selecting the larger eigenvectors as principal components implies information loss. There is no systematic way to determine which principal components (PCs) should be used. A new feature-extraction model to select the optimal principal components using ICSA is developed. The data acquired by the NASA airborne AVIRIS instrument over the Kennedy Space Center, Florida is used for evaluation. Experimental results show that our method can get better results.