Article
Chemistry, Analytical
Mahdiyeh Ghaffari, Nematollah Omidikia, Cyril Ruckebusch
Summary: A method is proposed for the joint selection of essential samples and variables in a data matrix for spectral unmixing, leading to a highly-reduced dataset with benefits such as minimized computational effort, meticulous data mining, easier model building, and better problem understanding or interpretation. The approach allows for reduction rates of over 99%, easy application of multivariate curve resolution - alternating least squares (MCR-ALS) on reduced data sets, and ready acquisition of full distribution maps and spectral profiles.
ANALYTICA CHIMICA ACTA
(2021)
Article
Geosciences, Multidisciplinary
Hormoz Izadi, Scott Keating
Summary: The Born series provides a framework for connecting earth model parameters and a wavefield characterized by perturbation around a known reference wavefield. The inverse scattering series provides both linear and nonlinear expressions for approximating earth model parameters. The linear expression accurately approximates the depth-varying velocity profile with low contrast, while the nonlinear expression is required for higher accuracy.
JOURNAL OF APPLIED GEOPHYSICS
(2023)
Article
Operations Research & Management Science
Wilfredo Sosa, Fernanda MP Raupp
Summary: In this article, we introduce the quadratic orthogonal projection problem and propose an iterative algorithm to solve it. Through testing with different quadratic functions, we demonstrate the algorithm's potential in areas such as computer graphics.
Article
Engineering, Electrical & Electronic
Haoyang Yu, Xiaodi Shang, Meiping Song, Jiaochan Hu, Tong Jiao, Qiandong Guo, Bing Zhang
Summary: This article introduces a novel spectral-spatial classification framework for hyperspectral images by combining collaborative representation and maximum margin projection. Experimental results demonstrate the effectiveness and practicality of the proposed methods for HSI classification tasks.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2021)
Article
Computer Science, Artificial Intelligence
Alice Porebski, Mohamed Alimoussa, Nicolas Vandenbroucke
Summary: This study explores texture analysis methods based on color and hyperspectral imaging and compares the applications of multi spectral band (MSB) and multi color channel (MCC) representations in texture classification. Experimental results show that considering interactions between components significantly improves the classification accuracy and the proposed approaches outperform state-of-the-art hand-designed and deep learning-based texture descriptors.
PATTERN RECOGNITION LETTERS
(2022)
Article
Engineering, Electrical & Electronic
Bing Yang, Hong Li, Ziyang Guo
Summary: This article proposes a novel deep similarity network (DSN) for hyperspectral image (HSI) classification, which improves classification accuracy by learning a new similarity measure of pixel pairs. By constructing a two-classification dataset, extracting features using two subnetworks, and computing similarity using a fusion subnetwork, the DSN demonstrates superiority in HSI classification.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2021)
Article
Geochemistry & Geophysics
Zeng Xiao-Niu, Li Xi-Hai, Yu Xiao-Tong, Liu Ji-Hao, Liu Dai-Zhi
Summary: This paper proposes a method for vertical derivative conversion based on convex sets projection, with the cutoff wavenumber of the filter determined by fitting a fractal model. The improved method shows better processing precision and efficiency compared to conventional methods.
APPLIED GEOPHYSICS
(2021)
Article
Geochemistry & Geophysics
Bin Pan, Qiaoying Qu, Xia Xu, Zhenwei Shi
Summary: In this article, a new structure-color preserving network (SCPNet) based on the joint attention mechanism is proposed for hyperspectral super-resolution (HSR). The SCPNet consists of three modules: structure-preserving module (SPM), color-preserving module (CPM), and cross-fusion module. Experimental results show that the proposed SCPNet has advantages on three benchmark datasets compared to state-of-the-art HSR methods.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Environmental Sciences
Lei Pan, Hengchao Li, Xiang Dai, Ying Cui, Xifeng Huang, Lican Dai
Summary: This paper proposes an unsupervised latent low-rank projection learning with graph regularization method for feature extraction and classification of hyperspectral images. By decomposing the latent low-rank matrix and applying graph regularization, discriminative features can be extracted and intrinsic subspace structures can be preserved, leading to improved performance. The use of local weighted average in a sliding window is also effective in further enhancing the performance.
Article
Computer Science, Artificial Intelligence
Chunlin He, Yong Zhang, Dunwei Gong, Xianfang Song, Xiaoyan Sun
Summary: In this article, an unsupervised multitask artificial bee colony (ABC) BS algorithm based on variable-size clustering (MBBS-VC) is proposed to simultaneously obtain multiple optimal band subsets with different sizes. Several new strategies are designed to improve the algorithm's performance, and experimental results verify the superiority of the proposed BS algorithm.
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION
(2022)
Article
Geochemistry & Geophysics
Shuzhen Zhang, Ting Lu, Wei Fu, Shutao Li
Summary: A novel superpixel-level hybrid discriminant analysis (SHDA) method is proposed for hyperspectral image (HSI) classification. The method takes advantage of superpixels in characterizing spatial-spectral shape-adaptive structure and the power of discriminant analysis in enhancing class-separability. Two specific discriminant analysis modules, superpixel-level local discriminant analysis (SLDA) and superpixel-level nonlocal discriminant analysis (SNDA), are designed to effectively capture the local and nonlocal spatial-spectral correlation information. The proposed SHDA method outperforms several state-of-the-art techniques on three real hyperspectral datasets.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Xiaofei Yang, Weijia Cao, Yao Lu, Yicong Zhou
Summary: This study proposes a hyperspectral image classification network that combines convolution neural networks with transformer structures, enabling the capture of subtle spectral differences and the conveyance of local spatial context information. Experimental results demonstrate that the proposed network outperforms existing transformers and state-of-the-art CNN-based methods on four benchmark datasets.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Yang-Jun Deng, Heng-Chao Li, Si-Qiao Tan, Junhui Hou, Qian Du, Antonio Plaza
Summary: This article proposes a t-linear tensor subspace learning (tLTSL) model for robust feature extraction of hyperspectral images (HSIs) based on t-product projection. The t-product projection, a newly defined tensor transformation method, maximally captures the intrinsic structure of tensor data. Through the integrated tensor low-rank and sparse decomposition, the model effectively removes noise corruption and maps the high-order hyperspectral data into a subspace with comprehensive information. Additionally, a proposition related to tensor rank is proved to interpret the meaning of the tLTSL model. Extensive experiments on HSI data corrupted by simulated and real noise validate the effectiveness of tLTSL.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Computer Science, Artificial Intelligence
Suhad Lateef Al-Khafaji, Jun Zhou, Xiao Bai, Yuntao Qian, Alan Wee-Chung Liew
Summary: In this paper, a novel method for boundary detection in close-range hyperspectral images is proposed. The method effectively predicts the boundaries of objects with similar color but different materials. By estimating the spatial distribution of spectral responses and using abundance maps and spectral feature vectors, the method constructs a boundary map. Experimental results show that the proposed method outperforms alternative methods when dealing with boundaries of objects with similar color but different materials.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2022)
Article
Environmental Sciences
Clara Cruz-Ramos, Beatriz P. Garcia-Salgado, Rogelio Reyes-Reyes, Volodymyr Ponomaryov, Sergiy Sadovnychiy
Summary: The proposed method reduces data dimension by extracting Gabor texture features and utilizing LDA, followed by classification using ANN to build a data matrix for analysis. It achieves high classification rates but with longer training time compared to non-reduced features.