高光谱遥感数据光谱特征提取算法与分类研究

针对高光谱数据的特点,探讨了高光谱数据特征提取的若干算法,重点研究了导数光谱和光谱编码技术,并从地物光谱曲线中提取了其光谱吸收特征.对同类曲线特征求交得到识别地物的有效特征;对不同类曲线特征求交得到区分不同类地物的有效特征.最后基于提取的特征建立了地物识别决策树,从而达到快速识别分类地物的目的,能够实现依据地物光谱特征的地物识别与分类.     

Relevance-based feature extraction for hyperspectral images.

Hyperspectral imagery affords researchers all discriminating details needed for fine delineation of many material classes. This delineation is essential for scientific research ranging from geologic to environmental impact studies. In a data mining scenario, one cannot blindly discard information because it can destroy discovery potential. In a supervised classification scenario, however, the preselection of classes presents one with an opportunity to extract a reduced set of meaningful features without degrading classification performance. Given the complex correlations found in hyperspectral data and the potentially large number of classes, meaningful feature extraction is a difficult task. We turn to the recent neural paradigm of generalized relevance learning vector quantization (GRLVQ) [B. Hammer and T. Villmann, Neural Networks vol. 15, pp. 1059-1068, 2002], which is based on, and substantially extends, learning vector quantization (LVQ) [T. Kohonen, Self-Organizing Maps, Berlin, Germany: Springer-Verlag, 2001] by learning relevant input dimensions while incorporating classification accuracy in the cost function. By addressing deficiencies in GRLVQ, we produce an improved version, GRLVQI, which is an effective analysis tool for high-dimensional data such as remotely sensed hyperspectral data. With an independent classifier, we show that the spectral features deemed relevant by our improved GRLVQI result in a better classification for a predefined set of surface materials than using all available spectral channels.     

高光谱图像空谱特征提取综述

由于高光谱图像包含了丰富的光谱、空间和辐射信息，且具有光谱接近连续、图谱合一的特性，可用于地质勘探、精细农业、生态环境、城市遥感以及军事目标检测等领域的目标精准分类与识别。对高光谱图像进行空谱特征提取是遥感领域的研究热点和前沿课题之一。传统空谱特征提取方法对高光谱图像分类的计算量和样本需求小、理论可解释性好、抗噪声能力强，但应用于分类的精度受限于特征来源；基于深度学习的高光谱图像空谱特征提取方法虽然计算量和样本需求大，但是由于深层空谱特征的表达能力更好，可以大幅度提高分类器的性能。为了便于对高光谱图像空谱特征提取领域进行更深入有效的探索，本文系统综述了相关研究进展。首先，概述了空间纹理与形态学特征提取、空间邻域信息获取及空间信息后处理等传统高光谱空谱特征提取方法的原理，对大量的已有工作进行了梳理、分析与总结。然后，从深度空谱特征提取角度出发，介绍了当前流行的卷积神经网络、图卷积神经网络及跨场景多源数据模型的结构特点及研究进展，分析、评价了基于深度学习的网络模型对高光谱图像空谱特征提取的优势及问题所在。最后，对该研究领域的未来相关发展提出建议并进行了展望。     

A new hybrid feature extraction method in a dyadic scheme for classification of hyperspectral data

The feature extraction is an important preprocessing step of the classification procedure particularly in high-dimensional data with limited number of training samples. Conventional supervised feature extraction methods, for example, linear discriminant analysis (LDA), generalized discriminant analysis, and non-parametric weighted feature extraction ones, need to calculate scatter matrices. In these methods, within-class and between-class scatter matrices are used to formulate the criterion of class separability. Because of the limited number of training samples, the accurate estimation of these matrices is not possible. So the classification accuracy of these methods falls in a small sample size situation. To cope with this problem, a new supervised feature extraction method namely, feature extraction using attraction points (FEUAP) has been recently proposed in which no statistical moments are used. Thus, it works well using limited training samples. To take advantage of this method and LDA one, this article combines them by a dyadic scheme. In the proposed scheme, the similar classes are grouped hierarchically by the k-means algorithm so that a tree with some nodes is constructed. Then the class of each pixel is determined from this scheme. To determine the class of each pixel, depending on the node of the tree, we use FEUAP or LDA for a limited or large number of training samples, respectively. The experimental results demonstrate the better performance of the proposed hybrid method in comparison with other supervised feature extraction methods in a small sample size situation.     

Application of the wavelet transform for feature extraction in the analysis of hyperspectral laser-induced fluorescence data

Remote-sensing approaches for environmental protection and exploration have evolved rapidly in the last decade. Among the new operational tools, hyperspectral Fluorescent LiDAR System (FLS®) lidar has demonstrated a high sensitivity and the ability to function in complex environments for real-time, robust oil-spill monitoring on airborne or ship-borne analytical platforms. The capabilities of such analytical platforms include real-time analysis of laser-induced fluorescence (LIF) data. Although numerous examples of the application of signal theory to the analysis of hyperspectral data appear in the remote-sensing literature, the conventional data analysis strategies are not well adapted to the practical issues of the LIF applications. The aim of this article is to provide a new approach for LIF lidar analytical platforms, which is focused on the specifics of hyperspectral LIF data. The approach is based on structural data analysis and interpretation, through which more detailed spectral matching is performed. This article is based on a simulated experiment in which the spectra of actual seawater and well-known types of petroleum products were combined to demonstrate the wavelet-transform-based analysis of LIF data. The final part of the article demonstrates the application of the wavelet transform to the structural analysis of LIF data from field experiments for the detection and identification of oil products in difficult environmental conditions.     

Effective feature extraction through segmentation-based folded-PCA for hyperspectral image classification

The remote sensing hyperspectral images (HSIs) usually comprise many important information of the land covers capturing through a set of hundreds of narrow and contiguous spectral wavelength bands. Appropriate classification performance can only offer the required knowledge from these immense bands of HSI since the classification result is not reasonable using all the original features (bands) of the HSI. Although it is not easy to calculate the intrinsic features from the bands, band (dimensionality) reduction techniques through feature extraction and feature selection are usually applied to increase the classification result and to fix the curse of dimensionality problem. Though the Principal Component Analysis (PCA) has been commonly adopted for the feature reduction of HSI, it can often fail to extract the local useful characteristics of the HSI for effective classification as it considers the global statistics of the HSI. Consequently, Segmented-PCA (SPCA), Spectrally-Segmented-PCA (SSPCA), Folded-PCA (FPCA) and Superpixelwise PCA (SuperPCA) have been introduced for better feature extraction of HSI in diverse ways. In this paper, feature extraction through SPCA & FPCA and SSPCA & FPCA, termed as Segmented-FPCA (SFPCA) and Spectrally-Segmented-FPCA (SSFPCA) respectively, has further been improved through applying FPCA on the highly correlated or spectrally separated bands’ segments of the HSI rather than not applying the FPCA on the entire dataset directly. The proposed methods are compared and analysed for a real mixed agricultural and an urban HSI classification using per-pixel SVM classifier. The experimental result shows that the classification performance using SSFPCA and SFPCA outperforms that of using conventional PCA, SPCA, SSPCA, FPCA, SuperPCA and using the entire original dataset without employing any feature reduction. Moreover, the proposed feature extraction methods provide the least memory and computation cost complexity.     

分段提取函数型数据特征的算法研究

针对函数型数据分类算法中全局统计特征表达能力有限,且显著点特征易受噪声干扰等问题,提出一种基于统计深度方法的函数曲线特征分段提取算法。首先,利用数据平滑技术对离散观测的数据进行平滑化处理,同时引入函数型数据的一阶和二阶导函数;然后,分段计算函数本身及其低阶导函数的马氏积分深度值,在此基础上构造函数曲线特征向量;最后,给出三种选择调节参数的搜索方案,并进行分类研究。在UCR数据集上的实验表明,与当前其他曲线特征提取算法相比,所提算法能有效提取函数曲线特征,提高分类的准确性。     

Local discriminant non-negative matrix factorization feature extraction for hyperspectral image classification

Non-negative matrix factorization (NMF) ignores both the local geometric structure of and the discriminative information contained in a data set. A manifold geometry-based NMF dimension reduction method called local discriminant NMF (LDNMF) is proposed in this paper. LDNMF preserves not only the non-negativity but also the local geometric structure and discriminative information of the data. The local geometric and discriminant structure of the data manifold can be characterized by a within-class graph and a between-class graph. An efficient multiplicative updating procedure is produced, and its global convergence is guaranteed theoretically. Experimental results on two hyperspectral image data sets show that the proposed LDNMF is a powerful and promising tool for extracting hyperspectral image features.     

Using class-based feature selection for the classification of hyperspectral data

The rapid advances in hyperspectral sensing technology have made it possible to collect remote-sensing data in hundreds of bands. However, the data-analysis methods that have been successfully applied to multispectral data are often limited in achieving satisfactory results for hyperspectral data. The major problem is the high dimensionality, which deteriorates the classification due to the Hughes Phenomenon. In order to avoid this problem, a large number of algorithms have been proposed, so far, for feature reduction. Based on the concept of multiple classifiers, we propose a new schema for the feature selection procedure. In this framework, instead of using feature selection for whole classes, we adopt feature selection for each class separately. Thus different subsets of features are selected at the first step. Once the feature subsets are selected, a Bayesian classifier is trained on each of these feature subsets. Finally, a combination mechanism is used to combine the outputs of these classifiers. Experiments are carried out on an Airborne Visible/Infrared Imaging Spectroradiometer (AVIRIS) data set. Encouraging results have been obtained in terms of classification accuracy, suggesting the effectiveness of the proposed algorithms.     

Self-adaptive differential evolution for feature selection in hyperspectral image data

Hyperspectral images are captured from hundreds of narrow and contiguous bands from the visible to infrared regions of electromagnetic spectrum. Each pixel of an image is represented by a vector where the components of the vector constitute the reflectance value of the surface for each of the bands. The length of the vector is equal to the number of bands. Due to the presence of large number of bands, classification of hyperspectral images becomes computation intensive. Moreover, higher correlation among neighboring bands increases the redundancy among them. As a result, feature selection becomes very essential for reducing the dimensionality. In the proposed work, an attempt has been made to develop a supervised feature selection technique guided by evolutionary algorithms. Self-adaptive differential evolution (SADE) is used for feature subset generation. Generated subsets are evaluated using a wrapper model where fuzzy k-nearest neighbor classifier is taken into consideration. Our proposed method also uses a feature ranking technique, ReliefF algorithm, for removing duplicate features. To demonstrate the effectiveness of the proposed method, investigation is carried out on three sets of data and the results are compared with four other evolutionary based state-of-the-art feature selection techniques. The proposed method shows promising results compared to others in terms of overall classification accuracy and Kappa coefficient.     

Moment-based feature extraction from high spatial resolution hyperspectral images

In this article, we propose a method for extracting spatio-spectral features from high spatial resolution hyperspectral (HS) images. The method is based on extracting two-dimensional moments from neighbourhoods of pixels. Three different types of moments are considered: geometric, complex Zernike and Legendre. Moments of a given type are extracted from a few principal components (PC) of HS data, and are stacked on the original HS data to form a joint spatio-spectral feature space. These features are classified using a support vector machine (SVM) classifier. The influence of the moments orders and the size of the neighbourhood window on the quality of the extracted features are analysed. A few experiments are conducted on two widely used HS data sets, Pavia University and Salinas. The results demonstrate high capabilities of the proposed method in comparison with some state-of-the-art spatio-spectral HS classification methods.     

Fast hyperspectral band selection based on spatial feature extraction

Hyperspectral images usually consist of hundreds of spectral bands, which can be used to precisely characterize different land cover types. However, the high dimensionality also has some disadvantages, such as the Hughes effect and a high storage demand. Band selection is an effective method to address these issues. However, most band selection algorithms are conducted with the high-dimensional band images, which will bring high computation complexity and may deteriorate the selection performance. In this paper, spatial feature extraction is used to reduce the dimensionality of band images and improve the band selection performance. The experiment results obtained on three real hyperspectral datasets confirmed that the spatial feature extraction-based approach exhibits more robust classification accuracy when compared with other methods. Besides, the proposed method can dramatically reduce the dimensionality of each band image, which makes it possible for band selection to be implemented in real time situations.     

Semi-supervised fuzzy neighborhood preserving analysis for feature extraction in hyperspectral remote sensing images

Neural Computing and Applications - Semi-supervised feature extraction methods are an important focus of interest in data mining and machine learning areas. These methods are improved methods based...     

Spectral characteristics and feature selection of hyperspectral remote sensing data

Hyperspectral remote sensing data with bandwidth of nanometre (nm) level have tens or even several hundreds of channels and contain abundant spectral information. Different channels have their own properties and show the spectral characteristics of various objects in image. Rational feature selection from the varieties of channels is very important for effective analysis and information extraction of hyperspectral data. This paper, taking Shunyi region of Beijing as a study area, comprehensively analysed the spectral characteristics of hyperspectral data. On the basis of analysing the information quantity of bands, correlation between different bands, spectral absorption characteristics of objects and object separability in bands, a fundamental method of optimum band selection and feature extraction from hyperspectral remote sensing data was proposed.     

Cluster analysis of genome-wide expression data for feature extraction

Bio-chip data that consists of high-dimensional attributes have more attributes than specimens. Thus, it is difficult to obtain covariance matrix from tens thousands of genes within a number of samples. Feature selection and extraction is critical to remove noisy features and reduce the dimensionality in microarray analysis. This study aims to fill the gap by developing a data mining framework with a proposed algorithm for cluster analysis of gene expression data, in which coefficient correlation is employed to arrange genes. Indeed, cluster analysis of microarray data can find coherent patterns of gene expression. The output is displayed as table list for convenient survey. We adopt the breast cancer microarray dataset to demonstrate practical viability of this approach.     

A fast algorithm for feature extraction of hyperspectral images using the first order statistics

Multimedia Tools and Applications - A new supervised feature extraction method appropriate for small sample size situations is proposed in this work. The proposed method is based on the first-order...     

Logistic regression-based pattern classifiers for symbolic interval data

This paper introduces different pattern classifiers for interval data based on the logistic regression methodology. Four approaches are considered. These approaches differ according to the way of representing the intervals. The first classifier considers that each interval is represented by the centres of the intervals and performs a classic logistic regression on the centers of the intervals. The second one assumes each interval as a pair of quantitative variables and performs a conjoint classic logistic regression on these variables. The third one considers that each interval is represented by its vertices and a classic logistic regression on the vertices of the intervals is applied. The last one assumes each interval as a pair of quantitative variables, performs two separate classic logistic regressions on these variables and combines the results in some appropriate way. Experiments with synthetic data sets and an application with a real interval data set demonstrate the usefulness of these classifiers.     

Supervised two-step feature extraction for structured representation of text data

Training data matrix used for classification of text documents to multiple categories is characterized by large number of dimensions while the number of manually classified training documents is relatively small. Thus the suitable dimensionality reduction techniques are required to be able to develop the classifier. The article describes two-step supervised feature extraction method that takes advantage of projections of terms into document and category spaces. We propose several enhancements that make the method more efficient and faster than it was presented in our former paper. We also introduce the adjustment score that enables to correct defected targets or helps to identify improper training examples that bias extracted features.