Decision fusion for hyperspectral image classification based on multiple features and locality-preserving analysis

Abstract

A novel fusion-classification system is proposed for hyperspectral image classification. Firstly, spectral derivatives are used to capture salient spectral features for different land-cover classes and a Gabor filter is applied to extract useful spatial features at neighbouring locations. Then, two locality-preserving dimensionality reduction methods are employed to reduce the dimensionality of data and preserve the local structure of neighbouring samples in the original image, derivative-feature and Gabor-feature domains. Finally, the classification results from Gaussian-mixture-model classifiers are fused by a decision-fusion approach. We have compared the proposed system to several traditional and state-of-the-art methods on two benchmark classification data sets. In both cases, our system achieved improved accuracy than the current best performing methods. Especially in the case of classification for the Indian Pines data set, the proposed DG-locality-preserving nonnegative matrix factorization (LPNMF) has 7% higher accuracy than support vector machine-Markov random field and LPNMF-G-MRF, and also has about 4% higher accuracy than Gabor-LPNMF. The proposed method has practical relevance as it shows potential to create smooth classification maps even for complex, spatially heterogeneous land-cover classes.

Decision Fusion at Pixel Level of Multi-Band Data for Land Cover Classification—A Review

“…Further, Ref. [19] investigates decision fusion based on multiple features and locality-preserving analysis with GMM and LOGP as a decision fusion scheme. Advancements in feature extraction were also studied in [20][21][22], which incorporated morphological profiles with two classifiers, namely SVM and random forest (RF); joint collaborative representation (JCR) and SVM models; and Gabor features, respectively.…”

A weighted multiple-feature fusion classifier for hyperspectral images with limited training samples

“…where x and y are the coordinate position, δ is the wavelength of Gabor function, θ represents the orientation, ϕ is the phase offset, σ is the deviation of Gaussian envelope, and the spatial aspect ratio is represented by γ (Ye, Bai, & Nian, 2017). f ¼ 1=δ is the spatial frequency of cosine factor.…”

Lithological classification by PCA-QPSO-LSSVM method with thermal infrared hyper-spectral data

“…A relatively complete method system has been formed 22 – 29 A series of studies have shown that hyperspectral remote sensing data can distinguish the undistinguishable substances in multispectral remote sensing data due to its rich spectral information, which can help obtain accurate classification results. Zhang and Li 30 proposed the determining reference spectra method based on SAM for lithological mapping with EO-1 hyperion hyperspectral data.…”