Novel Folded-PCA for Improved Feature Extraction and Data Reduction with Hyperspectral Imaging and SAR in Remote Sensing

As a widely used approach for feature extraction and data reduction, Principal Components Analysis (PCA) suffers from high computational cost, large memory requirement and low efficacy in dealing with large dimensional datasets such as Hyperspectral Imaging (HSI). To this end, a novel Folded-PCA is proposed, in which the spectral vector is folded into a matrix to allow the covariance matrix to be determined in a more efficient way. With this matrix-based representation, both global and local structures can be extracted to provide additional information for data classification. In addition, both the computational cost and the memory requirement have been significantly reduced. Using Support Vector Machine (SVM) for classification on two well-known HSI datasets and one Synthetic Aperture Radar (SAR) dataset in remote sensing applications, quantitative results are generated for objective evaluations. Comprehensive results have indicated that the proposed Folded-PCA approach not only outperforms the conventional PCA but also the baseline approach where the whole feature sets are used. Keywords—Folded Principal Component Analysis (F-PCA), feature extraction, data reduction, Hyperspectral Imaging (HSI), Support Vector Machine (SVM), remote sensing. Corresponding Author: Dr Jinchang Ren Hyperspectral Imaging Centre, Dept. of Electronic and Electrical Engineering University of Strathclyde Glasgow, G1 1XW United Kingdom Email: [email protected] Tel. +44-141-5482384