Unsupervised spectral sub-feature learning for hyperspectral image classification

Abstract

Spectral pixel classification is one of the principal techniques used in hyperspectral image (HSI) analysis. In this article, we propose an unsupervised feature learning method for classification of hyperspectral images. The proposed method learns a dictionary of subfeature basis representations from the spectral domain, which allows effective use of the correlated spectral data. The learned dictionary is then used in encoding convolutional samples from the hyperspectral input pixels to an expanded but sparse feature space. Expanded hyperspectral feature representations enable linear separation between object classes present in an image. To evaluate the proposed method, we performed experiments on several commonly used HSI data sets acquired at different locations and by different sensors. Our experimental results show that the proposed method outperforms other pixel-wise classification methods that make use of unsupervised feature extraction approaches. Additionally, even though our approach does not use any prior knowledge, or labelled training data to learn features, it yields either advantageous, or comparable, results in terms of classification accuracy with respect to recent semi-supervised methods. ARTICLE HISTORY

Spectral-Spatial Joint Classification of Hyperspectral Image Based on Broad Learning System

“…is the concatenation of all the first l groups of enhancement nodes [19]. Combined with Equation (4), the output result of BLS can be expressed by Equation (5) (5) where W op is the connecting weight matrix from all mapped feature nodes and all enhancement nodes to the output of the BLS. The superscript "op" represents the optimal weight [19].…”

“…The spectral feature extraction of HSI can be realized by unsupervised [5,6], supervised [7,8], and semi-supervised methods [7,9,10]. Representative unsupervised methods include principal component analysis (PCA) [11], independent component analysis (ICA) [12], and locality preserving projections (LPP) [13].…”

Hyperspectral Image Classification Based on 3D Convolution Joint Attention Mechanism

“…近年来，国内外高光谱遥感技术得到了快速发 展，因此在农业，环境科学，地物观测等方面取得了 广泛的应用 [1][2][3][4] 。高光谱图像(Hyperspectral Image， HSI)是一个三维立方体的图像数据，它是由二维数 字图像和一个光谱维度组成，包含几十甚至几百个连 续的光谱波段，可以利用这些光谱信息对图像进行分 类。常用的分类方法包括：K 最近邻(K Nearest Neighbor，K-NN) [5][6][7] ，极端学习机(Extreme Learning Machine，ELM) [8] 以及支持向量机(Support Vector Machine，SVM) [9][10][11] 等， K 最近邻是最简易的分类器， 这些都是充分依赖光谱信息进行分类，而不是利用像 素之间的空间信息。其中，极端学习机(ELM)是一 种有效的高光谱图像数据分类方法，对分类准确率的 提高是非常有效。支持向量机(SVM)是以数量不多 的样本为前提，通过探求网络模型的复杂程度和学习 能力之间的最好调和， 得到最佳的分类准确率。 另外， 目前还有一些常用的特征选择方法用于高光谱图像 分 类 ， 例 如 ： 主 成 分 分 析 ( Principal Component Analysis，PCA) [12][13] ，它可以使用频段选择来减少 噪声和一些不重要的特征，有效地保留最重要的一些 特 征 ， 有 助 于 高 光 谱 图 像 分 类 。 独 立 成 分 分 析 (Independent Component Analysis，ICA) [14] 是一种利 用统计原理进行计算的方法，具有运算复杂程度低的 特点。 近些年基于深度学习的图像分类方法 [15][16][17][18] 受到 普遍的关注。文献 [19] [7] , SVM [9] , PCA [13] , ICA [14] , RNN [21] , 3D-CNN [20] 和 CNN-PPF [19]…”

A Hyperspectral Image Classification Framework with Spatial Pixel Pair Features

“…Additionally, the vast variations in the statistical distributions of channels/wavelengths also draw much research attention. Slavkovikj et al [ 45 ] propose an unsupervised sub-feature learning method in the spectral domain. This dictionary learning-based method greatly enhanced the hyperspectral feature representations.…”