scholarly journals Spectral-Similarity-Based Kernel of SVM for Hyperspectral Image Classification

2020 ◽  
Vol 12 (13) ◽  
pp. 2154 ◽  
Author(s):  
Ke Wang ◽  
Ligang Cheng ◽  
Bin Yong
Keyword(s):  
Hyperspectral Image ◽  
Similarity Measures ◽  
Kernel Functions ◽  
Spectral Information ◽  
Spectral Angle Mapper ◽  
Spectral Similarity ◽  
Information Divergence ◽  
Rbf Kernel ◽  
Hyperspectral Classification ◽  
Spectral Angle

Spectral similarity measures can be regarded as potential metrics for kernel functions, and can be used to generate spectral-similarity-based kernels. However, spectral-similarity-based kernels have not received significant attention from researchers. In this paper, we propose two novel spectral-similarity-based kernels based on spectral angle mapper (SAM) and spectral information divergence (SID) combined with the radial basis function (RBF) kernel: Power spectral angle mapper RBF (Power-SAM-RBF) and normalized spectral information divergence-based RBF (Normalized-SID-RBF) kernels. First, we prove these spectral-similarity-based kernels to be Mercer’s kernels. Second, we analyze their efficiency in terms of local and global kernels. Finally, we consider three hyperspectral datasets to analyze the effectiveness of the proposed spectral-similarity-based kernels. Experimental results demonstrate that the Power-SAM-RBF and SAM-RBF kernels can obtain an impressive performance, particularly the Power-SAM-RBF kernel. For example, when the ratio of the training set is 20 % , the kappa coefficient of Power-SAM-RBF kernel (0.8561) is 1.61 % , 1.32 % , and 1.23 % higher than that of the RBF kernel on the Indian Pines, University of Pavia, and Salinas Valley datasets, respectively. We present three conclusions. First, the superiority of the Power-SAM-RBF kernel compared to other kernels is evident. Second, the Power-SAM-RBF kernel can provide an outstanding performance when the similarity between spectral signatures in the same hyperspectral dataset is either extremely high or extremely low. Third, the Power-SAM-RBF kernel provides even greater benefits compared to other commonly used kernels when the sizes of the training sets increase. In future work, multiple kernels combining with the spectral-similarity-based kernel are expected to be provide better hyperspectral classification.

scholarly journals Cartografía del aguacate en el sur del estado de México mediante tratamiento digital de imágenes sentinel-2

2020 ◽  
Vol 11 (4) ◽  
pp. 865-879
Author(s):  
Dulce Karen Figueroa-Figueroa ◽  
Jose Francisco Ramírez Dávila ◽  
Xanat Antonio-Némiga ◽  
Andrés González Huerta
Keyword(s):  
Persea Americana ◽  
Spectral Information ◽  
Spectral Angle Mapper ◽  
Information Divergence ◽  
Spectral Angle ◽  
Sentinel 2

El cultivo de aguacate (Persea americana Mill.) es uno de los más importantes en México, entre los estados con mayor producción se encuentra el Estado de México, que es el tercer estado productor a nivel nacional. Coatepec Harinas y Donato Guerra son dos de los municipios más representativos en lo respectivo a esta actividad; sin embargo, no existe un censo que especifique la superficie del cultivo, por lo que el objetivo de esta investigación fue probar métodos de índices de vegetación, algoritmos spectral angle mapper (SAM) y spectral information divergence (SID) y la combinación de estos en las imágenes del sensor Sentinel-2 para evaluar su desempeño en la identificación de áreas plantadas con el cultivo de aguacate. Los resultados se validaron con una matriz de confusión y la comparación de los datos de referencia de entrenamiento y validación. El algoritmo SID alcanzó una precisión de 97.5% para detectar aguacate, mientras que el tratamiento SAM obtuvo una precisión de 63.1%. La combinación de SID con el índice Anthocyanin Reflectance Index 1 (ARI1), proporcionó un mejor resultado sobre la cartografía de validación regional con 85% de precisión. Otras combinaciones de índices y tratamientos dieron resultados inferiores al 50% de la precisión por lo que no se recomiendan. Esta metodología podría ser probada para la detección de otros cultivos de interés comercial, dado que Sentinel-2 muestra ser una alternativa viable para este tipo de estudios, teniendo una buena resolución espectral, además de ser de fácil acceso y manipulación.

scholarly journals AVALIAÇÃO DA ACURÁCIA DE ALGORITMOS DE CLASSIFICAÇÃO DE IMAGENS ORBITAIS NA BAÍA DA BABITONGA, NORDESTE DE SANTA CATARINA

2021 ◽  
Vol 14 (6) ◽  
pp. 3577
Author(s):  
Celso Voos Vieira ◽  
Pedro Apolonid Viana
Keyword(s):  
Feature Extraction ◽  
Maximum Likelihood ◽  
Mahalanobis Distance ◽  
Minimum Distance ◽  
Spectral Information ◽  
Classification Algorithms ◽  
Neural Net ◽  
Spectral Angle Mapper ◽  
Information Divergence ◽  
Spectral Angle

O objetivo deste trabalho foi a avaliação da acurácia de algoritmos de classificação do uso e cobertura do solo, quando aplicados a uma imagem orbital de média resolução espacial. Para esse estudo foram utilizadas as bandas espectrais da faixa do visível e infravermelho próximo, do sensor Operational Land Imager – OLI na Baía da Babitonga/SC. Foram propostas nove classes de cobertura do solo, que serviram como controle para testar 11 algoritmos classificadores: Binary Encoding, Example Based Feature Extraction, IsoData, K-Means, Mahalanobis Distance, Maximum Likelihood, Minimum Distance, Neural Net, Parallelepiped, Spectral Angle Mapper e Spectral Information Divergence. O classificador Maximum Likelihood foi o que apresentou o melhor desempenho, obtendo um índice Kappa de 0,89 e acurácia global de 95,5%, sendo capaz de distinguir as nove classes de cobertura do solo propostas. Evaluation of the Accuracy of Orbital Image Classification Algorithms in Babitonga Bay, northeast of Santa Catarina A B S T R A C TThe objective of this work was to evaluate the classification algorithms accuracy of the soil use and cover when applied to a spatial mean orbital image. For this study we used the visible and near infrared spectral bands of the Operational Land Imager - OLI sensor in Babitonga Bay / SC. Nine classes of soil cover were proposed, which served as control to test 11 classifier algorithms: Binary Encoding, Example Based Feature Extraction, IsoData, K-Means, Mahalanobis Distance, Maximum Likelihood, Minimum Distance, Neural Net, Parallelepiped, Spectral Angle Mapper and Spectral Information Divergence. The Maximum Likelihood classifier presented the best performance, obtaining a Kappa index of 0.89 and a global accuracy of 95.5%, being able to distinguish the nine proposed classes of soil cover.Keywords: Algorithms Accuracy, Babitonga Bay, Orbital image, Remote sensing, Soil Use and Cover. 

scholarly journals AN IMPROVED VARIATIONAL METHOD FOR HYPERSPECTRAL IMAGE PANSHARPENING WITH THE CONSTRAINT OF SPECTRAL DIFFERENCE MINIMIZATION

2017 ◽  
Vol XLII-2/W7 ◽  
pp. 753-760
Author(s):  
Z. Huang ◽  
Q. Chen ◽  
Y. Shen ◽  
Q. Chen ◽  
X. Liu
Keyword(s):  
Variational Method ◽  
Energy Function ◽  
Hyperspectral Image ◽  
Spectral Information ◽  
Gradient Descent Method ◽  
Equal Proportion ◽  
Spectral Distortion ◽  
Spectral Difference ◽  
Spectral Angle Mapper ◽  
Fidelity Term

Variational pansharpening can enhance the spatial resolution of a hyperspectral (HS) image using a high-resolution panchromatic (PAN) image. However, this technology may lead to spectral distortion that obviously affect the accuracy of data analysis. In this article, we propose an improved variational method for HS image pansharpening with the constraint of spectral difference minimization. We extend the energy function of the classic variational pansharpening method by adding a new spectral fidelity term. This fidelity term is designed following the definition of spectral angle mapper, which means that for every pixel, the spectral difference value of any two bands in the HS image is in equal proportion to that of the two corresponding bands in the pansharpened image. Gradient descent method is adopted to find the optimal solution of the modified energy function, and the pansharpened image can be reconstructed. Experimental results demonstrate that the constraint of spectral difference minimization is able to preserve the original spectral information well in HS images, and reduce the spectral distortion effectively. Compared to original variational method, our method performs better in both visual and quantitative evaluation, and achieves a good trade-off between spatial and spectral information.

scholarly journals Spatial Prior Fuzziness Pool-Based Interactive Classification of Hyperspectral Images

2019 ◽  
Vol 11 (9) ◽  
pp. 1136 ◽  
Author(s):  
Muhammad Ahmad ◽  
Asad Khan ◽  
Adil Mehmood Khan ◽  
Manuel Mazzara ◽  
Salvatore Distefano ◽  
...  
Keyword(s):  
Hyperspectral Image ◽  
Sample Selection ◽  
Spectral Angle Mapper ◽  
Classifier Design ◽  
Heterogeneous Samples ◽  
Marginal Probability Distribution ◽  
Benchmark Datasets ◽  
Manual Selection ◽  
Spectral Angle

Acquisition of labeled data for supervised Hyperspectral Image (HSI) classification is expensive in terms of both time and costs. Moreover, manual selection and labeling are often subjective and tend to induce redundancy into the classifier. Active learning (AL) can be a suitable approach for HSI classification as it integrates data acquisition to the classifier design by ranking the unlabeled data to provide advice for the next query that has the highest training utility. However, multiclass AL techniques tend to include redundant samples into the classifier to some extent. This paper addresses such a problem by introducing an AL pipeline which preserves the most representative and spatially heterogeneous samples. The adopted strategy for sample selection utilizes fuzziness to assess the mapping between actual output and the approximated a-posteriori probabilities, computed by a marginal probability distribution based on discriminative random fields. The samples selected in each iteration are then provided to the spectral angle mapper-based objective function to reduce the inter-class redundancy. Experiments on five HSI benchmark datasets confirmed that the proposed Fuzziness and Spectral Angle Mapper (FSAM)-AL pipeline presents competitive results compared to the state-of-the-art sample selection techniques, leading to lower computational requirements.

Sign in / Sign up

Export Citation Format

Share Document