Classification of hyperspectral images with regularized linear discriminant analysis (original) (raw)

Folded LDA: Extending the Linear Discriminant Analysis Algorithm for Feature Extraction and Data Reduction in Hyperspectral Remote Sensing

IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING, 2021

The rich spectral information provided by hyperspectral imaging has made this technology very useful in the classification of remotely sensed data. However, classification of hyperspectral data is typically affected by noise and the Hughes phenomenon due to the presence of hundreds of spectral bands and correlation among them, with usually a limited number of samples for training. Linear discriminant analysis (LDA) is a well-known technique that has been widely used for supervised dimensionality reduction of hyperspectral data. However, the use of LDA in hyperspectral remote sensing is limited due to its poor performance on small training datasets and the limited number of features that can be selected i.e., c − 1, where c is the number of classes in the data. To solve these problems, this article presents a folded LDA (F-LDA) for dimensionality reduction of remotely sensed HSI data in small sample size scenarios. The proposed approach allows many more discriminant features to be selected in comparison to the conventional LDA since the selection is no longer bound by the limiting factor, leading to significantly higher accuracy in the classification of pixels under SSS restrictions. The proposed approach is evaluated on five different datasets, where the experimental results demonstrate the superiority of the F-LDA to the conventional LDA in terms of not only higher classification accuracy but also reduced computational complexity, and reduced contiguous memory requirements.

Robust Hyperspectral Image Classification by Multi-Layer Spatial-Spectral Sparse Representations

Sparse representation (SR)-driven classifiers have been widely adopted for hyperspectral image (HSI) classification, and many algorithms have been presented recently. However, most of the existing methods exploit the single layer hard assignment based on class-wise reconstruction errors on the subspace assumption; moreover, the single-layer SR is biased and less stable due to the high coherence of the training samples. In this paper, motivated by category sparsity, a novel multi-layer spatial-spectral sparse representation (mlSR) framework for HSI classification is proposed. The mlSR assignment framework effectively classifies the test samples based on the adaptive dictionary assembling in a multi-layer manner and intrinsic class-dependent distribution. In the proposed framework, three algorithms, multi-layer SR classification (mlSRC), multi-layer collaborative representation classification (mlCRC) and multi-layer elastic net representation-based classification (mlENRC) for HSI, are developed. All three algorithms can achieve a better SR for the test samples, which benefits HSI classification. Experiments are conducted on three real HSI image datasets. Compared with several state-of-the-art approaches, the increases of overall accuracy (OA), kappa and average accuracy (AA) on the Indian Pines image range from 3.02% to 17.13%, 0.034 to 0.178 and 1.51% to 11.56%, respectively. The improvements in OA, kappa and AA for the University of Pavia are from 1.4% to 21.93%, 0.016 to 0.251 and 0.12% to 22.49%, respectively. Furthermore, the OA, kappa and AA for the Salinas image can be improved from 2.35% to 6.91%, 0.026 to 0.074 and 0.88% to 5.19%, respectively. This demonstrates that the proposed mlSR framework can achieve comparable or better performance than the state-of-the-art classification methods.