Semisupervised Sparse Manifold Discriminative Analysis for Feature Extraction of Hyperspectral Images | IEEE Journals & Magazine | IEEE Xplore

Semisupervised Sparse Manifold Discriminative Analysis for Feature Extraction of Hyperspectral Images


Abstract:

The graph embedding (GE) framework is very useful to extract the discriminative features of hyperspectral images (HSIs) for classification. However, a major challenge of ...Show More

Abstract:

The graph embedding (GE) framework is very useful to extract the discriminative features of hyperspectral images (HSIs) for classification. However, a major challenge of GE is how to select a proper neighborhood size for graph construction. To overcome this drawback, a new semisupervised discriminative learning algorithm, which is called the semisupervised sparse manifold discriminative analysis (S3MDA) method, was proposed by using manifold-based sparse representation (MSR) and GE. The proposed algorithm utilizes MSR to obtain the sparse coefficients of labeled and unlabeled samples. Then, it constructs a within-class graph and a between-class graph using the sparse coefficients of labeled samples, as well as an unsupervised graph with the sparse coefficients of unlabeled samples. Finally, it uses these graphs to obtain a projection matrix for feature extraction (FE) of HSI in a low-dimensional space. The S3MDA method not only inherits the merits of MSR to reveal the sparse manifold properties of data but also enhances interclass separability and intraclass compactness to improve the discriminating power for classification. Extensive experiments on two real HSI data sets obtained with a reflective optics system imaging spectrometer and an airborne visible/infrared imaging spectrometer show that the proposed algorithm is significantly superior to other state-of-the-art FE methods in terms of classification accuracy.
Published in: IEEE Transactions on Geoscience and Remote Sensing ( Volume: 54, Issue: 10, October 2016)
Page(s): 6197 - 6211
Date of Publication: 11 July 2016

ISSN Information:

Funding Agency:


I. Introduction

A hyperspectral image (HSI) contains dozens or even hundreds of contiguous spectral bands covering the electromagnetic spectrum from visible to near-infrared regions [1], [2]. Such images are captured by hyperspectral sensors [3]. In recent years, HSIs have been widely used in the fields of environmental monitoring, precision agriculture, object recognition, and land cover classification [4]–[6]. Despite their high-dimensional characteristic, HSI data are typically redundant with underlying structures that can be represented by only a few features [7], [8]. Therefore, the main challenge for HSI classification is to reduce the dimensionality of data points and remove redundancy. Applying feature reduction, including feature selection (FS) and feature extraction (FE), is an effective way to overcome this challenge and preserve relevant information in data points [9], [10]. FS involves directly choosing the best subbands from all spectral bands on the basis of certain criteria, whereas FE can be used to map high-dimensional data into a low-dimensional feature space with a number of FE algorithms. In this paper, we only focus on the FE method.

References

References is not available for this document.