Applying class-based feature extraction approaches for supervised classification of hyperspectral imagery
نویسندگان
چکیده
Global band selection or feature extraction methods have been applied to hyperspectral image classification to overcome the “curse of dimension”. We applied class-based feature extraction approaches and compressed the class data into different lower dimensional subspaces. Land cover classes in hyperspectral imagery could be roughly modelled as lowdimensional Gaussian clusters (i.e., “Gaussian pancakes”) floating in sparse hyperspace. Each pixel was labelled accordingly based on conventional classifiers. We evaluated and compared the class-based version of principal components analysis (PCA), probabilistic principal components analysis (PPCA), and probabilistic factor analysis (PFA) algorithms to find the lower dimensional class subspaces in the training stage, projected each pixel, and then assigned the class label according to the maximum likelihood decision rule. Results from simulations and the classification of a compact airborne spectrographic imager 2 (CASI 2) hyperspectral dataset were presented. The proposed class-based PCA (CPCA) algorithm provided a reasonable trade-off between classification accuracy and computational efficiency for hyperspectral image classification. It proved more efficient and provided the highest classification kappa coefficient (0.946) among all band selection and feature extraction classifiers in our study. CPCA is recommended as a useful class-based feature extraction method for classification of hyperspectral imagery. Résumé. Les méthodes de sélection de bandes ou d’extraction des caractéristiques ont été appliquées à la classification d’images hyperspectrales pour remédier au problème du fléau de la dimension. Nous avons appliqué des approches d’extraction des caractéristiques basées sur la classe et compressé les données de classes en différents sous-espaces de dimension plus faible. Les classes de couvert dans les images hyperspectrales peuvent être modélisées en gros comme des regroupements gaussiens de faible dimension (c.-à-d. « Gaussian pancakes ») flottant dans l’hyperespace. Chaque pixel a été étiqueté ainsi basé sur des classifieurs conventionnels. Nous avons évalué et comparé la version basée sur la classe des algorithmes d’analyse en composantes principales (ACP), d’analyse en composantes principales probabiliste (ACPP) et d’analyse factorielle probabiliste (AFP) pour trouver les sous-espaces de classes de plus petite dimension dans la phase d’entraînement, puis projeté chaque pixel et ensuite assigné l’étiquette de classe selon la règle de décision basée sur le maximum de vraisemblance. Les résultats des simulations et de la classification d’un ensemble de données hyperspectrales du capteur CASI 2 (« compact airborne spectrographic imager 2 ») sont présentés. L’algorithme ACP basé sur la classe (CPCA) constitue un compromis raisonnable entre la précision de classification et l’efficacité de calcul pour la classification d’images hyperspectrales. Il s’est avéré plus efficace et a donné le coefficient de classification kappa le plus élevé (0,946) parmi tous les classifieurs par sélection de bandes et d’extraction des caractéristiques dans notre étude. L’algorithme CPCA est recommandé comme méthode d’extraction des caractéristiques basée sur la classe pour la classification des images hyperspectrales. [Traduit par la Rédaction]
منابع مشابه
Overlap-based feature weighting: The feature extraction of Hyperspectral remote sensing imagery
Hyperspectral sensors provide a large number of spectral bands. This massive and complex data structure of hyperspectral images presents a challenge to traditional data processing techniques. Therefore, reducing the dimensionality of hyperspectral images without losing important information is a very important issue for the remote sensing community. We propose to use overlap-based feature weigh...
متن کاملHyperspectral Image Classification Based on the Fusion of the Features Generated by Sparse Representation Methods, Linear and Non-linear Transformations
The ability of recording the high resolution spectral signature of earth surface would be the most important feature of hyperspectral sensors. On the other hand, classification of hyperspectral imagery is known as one of the methods to extracting information from these remote sensing data sources. Despite the high potential of hyperspectral images in the information content point of view, there...
متن کاملکاهش ابعاد دادههای ابرطیفی به منظور افزایش جداییپذیری کلاسها و حفظ ساختار داده
Hyperspectral imaging with gathering hundreds spectral bands from the surface of the Earth allows us to separate materials with similar spectrum. Hyperspectral images can be used in many applications such as land chemical and physical parameter estimation, classification, target detection, unmixing, and so on. Among these applications, classification is especially interested. A hyperspectral im...
متن کاملSupervised Feature Extraction of Face Images for Improvement of Recognition Accuracy
Dimensionality reduction methods transform or select a low dimensional feature space to efficiently represent the original high dimensional feature space of data. Feature reduction techniques are an important step in many pattern recognition problems in different fields especially in analyzing of high dimensional data. Hyperspectral images are acquired by remote sensors and human face images ar...
متن کاملFeature reduction of hyperspectral images: Discriminant analysis and the first principal component
When the number of training samples is limited, feature reduction plays an important role in classification of hyperspectral images. In this paper, we propose a supervised feature extraction method based on discriminant analysis (DA) which uses the first principal component (PC1) to weight the scatter matrices. The proposed method, called DA-PC1, copes with the small sample size problem and has...
متن کاملانجام یک مرحله پیش پردازش قبل از مرحله استخراج ویژگی در طبقه بندی داده های تصاویر ابر طیفی
Hyperspectral data potentially contain more information than multispectral data because of their higher spectral resolution. However, the stochastic data analysis approaches that have been successfully applied to multispectral data are not as effective for hyperspectral data as well. Various investigations indicate that the key problem that causes poor performance in the stochastic approaches t...
متن کامل