Article
Agriculture, Multidisciplinary
Gangshan Wu, Yinlong Fang, Qiyou Jiang, Ming Cui, Na Li, Yunmeng Ou, Zhihua Diao, Baohua Zhang
Summary: This study used hyperspectral imaging combined with spectral features, vegetation indices, and textural features to detect gray mold on strawberry leaves. The results showed that the models based on optimum wavelengths and significant vegetation indices performed well, with a maximum classification accuracy of 93.33%. The models with combined features outperformed the models based on single features, with an accuracy range of 93.33-96.67%. Overall, the combined feature-based method significantly improved the recognition accuracy of strawberry gray mold and accurately identified infected leaves in the early stages.
COMPUTERS AND ELECTRONICS IN AGRICULTURE
(2023)
Article
Computer Science, Artificial Intelligence
Jun Wang, Chang Tang, Zhenglai Li, Xinwang Liu, Wei Zhang, En Zhu, Lizhe Wang
Summary: A novel band selection method based on region-aware latent features fusion was proposed, which utilized superpixel segmentation and Laplacian matrix construction to preserve spatial information and enhance separability among bands in hyperspectral images. The method achieved superior performance in comparison with other state-of-the-art methods through k-means clustering algorithm to obtain the index of selected bands.
INFORMATION FUSION
(2022)
Article
Geochemistry & Geophysics
Bo Yang, Yi He, Changzhe Jiao, Xiao Pan, Guozhen Wang, Lei Wang, Jinjian Wu
Summary: This article proposes a multiple-instance metric learning neural network (MIML-Net) for hyperspectral target detection tasks, which only requires region-level labels and greatly alleviates the laborious pixel-level annotation problems.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Environmental Sciences
Zhongheng Li, Fang He, Haojie Hu, Fei Wang, Weizhong Yu
Summary: This study proposed a novel anomaly detection approach, named Random Collective Representation-based Detector with Multiple Feature (RCRDMF), which improves the accuracy of hyperspectral anomaly detection (HAD) by combining multiple spectral and spatial features. Experimental results show the superiority of the proposed approach in terms of accuracy and speed.
Article
Engineering, Electrical & Electronic
Longshan Yang, Junhuan Peng, Yuebin Wang, Linlin Xu, Weiwei Zhu
Summary: In this article, a superpixel-based graph model is proposed for hyperspectral image classification. By utilizing anchor graph regularization, the model is able to extract local and nonlocal spatial information effectively. The approach constructs a scalable anchor graph using a small number of anchor points, which outperforms other semisupervised HSI classification approaches in cases with a limited quantity of labeled samples.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Yue Zhao, Jiangtao Peng, Yantao Wei, Qinmu Peng, Yi Mou
Summary: This letter proposes a method based on multiple-feature latent space learning for hyperspectral image classification, which aims to improve classification performance by learning the latent space and transformation matrices of multiple features, and using spatial information to label unlabeled samples. Experimental results on the Indian Pines and University of Pavia datasets demonstrate the effectiveness of the proposed method.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2021)
Article
Environmental Sciences
Zeyu Xu, Cheng Su, Shirou Wang, Xiaocan Zhang
Summary: In this study, a local-global spectral feature (LGSF) extraction and optimization method is proposed for hyperspectral image (HSI) classification. The method transforms the 1D spectral vector into a 2D spectral image and automatically extracts the LGSF by using the local spectral feature extraction module (LSFEM) and the global spectral feature extraction module (GSFEM). The LGSF is further optimized using a loss function inspired by contrastive learning. The proposed method demonstrates its effectiveness in utilizing spectral information and achieving accurate HSI classification.
Article
Geochemistry & Geophysics
Xiaojun Yang, Xiaobei Huang, Mingjun Zhu, Sha Xu, Yijun Liu
Summary: This paper proposes an ensemble and random RX anomaly detector with multiple features, which extracts multiple features and fuses detection results using ensemble learning method, achieving more accurate hyperspectral anomaly detection.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Biochemistry & Molecular Biology
Andi Nur Nilamyani, Firda Nurul Auliah, Mohammad Ali Moni, Watshara Shoombuatong, Md Mehedi Hasan, Hiroyuki Kurata
Summary: Nitrotyrosine, a type of protein post-translational modification, is generated by reactive nitrogen species. Computational prediction, such as the PredNTS predictor developed in this study, plays a vital role in understanding nitrated proteins before biological experimentation. The PredNTS predictor outperforms existing predictors and provides a useful computational resource for predicting nitrotyrosine sites.
INTERNATIONAL JOURNAL OF MOLECULAR SCIENCES
(2021)
Article
Chemistry, Multidisciplinary
Ling Dai, Guangyun Zhang, Jinqi Gong, Rongting Zhang
Summary: This paper proposes a data-driven method for hyperspectral remotely sensed data, which can autonomously extract key features and interactively learn feature indexes, providing a more flexible and creative framework compared to traditional methods.
APPLIED SCIENCES-BASEL
(2021)
Article
Optics
Xiaoyue Wang, Junyi Nan, Jiayun Xue, Weiwei Liu, Ming Yan, Shuai Yuan, Kun Huang, Heping Zeng
Summary: This study reports a simple technique based on filament-induced nonlinear spectroscopy for information-rich gas imaging. The technique enables high-spatial resolution point-wise imaging in a non-cooperative environment without the need for meticulously arranged cameras.
OPTICS AND LASERS IN ENGINEERING
(2022)
Article
Geochemistry & Geophysics
Yule Duan, Hong Huang, Tao Wang
Summary: The article introduces a new method called GSMH, which is a geodesic-based sparse manifold hypergraph, to address the small sample problem in HSI data. This method utilizes geodesic distance and a geodesic-based neighborhood SR model to explore sparse correlations among different manifold neighborhoods, then constructs a pair of semisupervised hypergraphs to obtain nonlinear discriminative feature representation.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Computer Science, Artificial Intelligence
Yingying Chen, Xiaowei Yang
Summary: In the field of support vector machines, online random feature map algorithms play a crucial role in large-scale nonlinear classification problems. However, the existing methods have shortcomings that can be overcome by the proposed random features based online adaptive kernel learning (RF-OAK) method, which shows superior performance through theoretical analysis and experiments.
PATTERN RECOGNITION
(2022)
Article
Environmental Sciences
Haojin Tang, Yanshan Li, Zhiquan Huang, Li Zhang, Weixin Xie
Summary: This paper proposes a fusion framework for small-sample hyperspectral image (HSI) classification. It extracts handcrafted spatial-spectral features using a 3D fuzzy histogram of oriented gradients (3D-FHOG) descriptor and combines CNN-based spatial-spectral features using a multidimensional Siamese network (MDSN). Experimental results show that the proposed fusion framework outperforms representative handcrafted feature-based and CNN-based methods.
Article
Chemistry, Multidisciplinary
Long Chen, Rining Wu, Feixiang Zhou, Huifeng Zhang, Jian K. Liu
Summary: This study introduces HybridGCN, a pioneering Hybrid Graph Convolutional Network that elevates solubility prediction accuracy through the combination of diverse features, encompassing sophisticated deep-learning features and classical biophysical features. Exploration into the interplay between deep-learning features and biophysical features revealed that specific biophysical attributes complement features extracted by advanced deep-learning models. The use of ESM, a substantial protein language model, enhanced the model's capability for feature representation, while the proposed Adaptive Feature Re-weighting (AFR) module enabled fine-tuning of feature importance.
JOURNAL OF CHEMINFORMATICS
(2023)
Article
Environmental Sciences
Lianhui Liang, Shaoquan Zhang, Jun Li, Antonio Plaza, Zhi Cui
Summary: We propose a new HSI classification approach, called MOCNN, which combines 2D octave and 3D CNNs to obtain complex spatial contextual feature information and spectral characteristics. Our approach outperforms several other methods for HSI classification, especially in scenarios dominated by limited and imbalanced sample data, through the use of attention mechanisms and sample balancing strategies.
Article
Environmental Sciences
Yunchang Wang, Jiang Cai, Junlong Zhou, Jin Sun, Yang Xu, Yi Zhang, Zhihui Wei, Javier Plaza, Antonio Plaza, Zebin Wu
Summary: This article introduces a new real-time anomaly detection algorithm called CE-RX for hyperspectral images. The algorithm combines the advantages of cloud and edge computing. The experimental results show that this algorithm can obtain more accurate results than existing real-time detection algorithms.
Article
Automation & Control Systems
Yuanchao Su, Lianru Gao, Mengying Jiang, Antonio Plaza, Xu Sun, Bing Zhang
Summary: This paper proposes a semi-supervised hyperspectral image classification method based on normalized spectral clustering and kernel-based learning. By aggregating local-to-global correlations, distinguishable features are extracted and used for classification along with a kernel-based extreme learning machine. Experimental results demonstrate the competitive performance of the proposed method on several hyperspectral images.
IEEE TRANSACTIONS ON CYBERNETICS
(2023)
Article
Geochemistry & Geophysics
Zhiyong Lv, Pengfei Zhang, Weiwei Sun, Jon Atli Benediktsson, Junhuai Li, Wei Wang
Summary: In this article, two novel features, the Gaussian-weighting spectral (GWS) feature and the area shape index (ASI) feature, are proposed to improve land cover classification with high spatial resolution remotely sensed imagery. Experimental results show that the proposed features can enhance classification accuracies and complement each other to improve classification performance.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Geochemistry & Geophysics
Liguo Wang, Xiaoyi Wang, Anna Vizziello, Paolo Gamba
Summary: Autoencoder is widely used in hyperspectral anomaly detection. We propose a residual self-attention-based AE (RSAAE) to distinguish between background and anomalies. Through the design of novel residual self-attention modules and a low-rank loss function, RSAAE outperforms eight popular methods in terms of detection accuracy on four real hyperspectral image datasets.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Geochemistry & Geophysics
Elena C. Rodriguez-Garlito, Abel Paz-Gallardo, Antonio J. Plaza
Summary: Multispectral remote sensing is effective in detecting different land covers, including invasive aquatic plants. This study presents a new methodology using satellite images to automatically identify the areas with frequent accumulation of invasive aquatic plants in the Guadiana River. Deep learning (CNN) is applied for plant detection, followed by GIS analysis to map the locations of water hyacinth patches. The successful management of invasive aquatic plants in the Guadiana River is demonstrated.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2023)
Article
Computer Science, Artificial Intelligence
Chengle Zhou, Qian Shi, Da He, Bing Tu, Haoyang Li, Antonio Plaza
Summary: This article introduces a novel convolutional transformer method based on spectral-spatial sequence characteristics for ground coverings change detection using bitemporal hyperspectral images. The proposed method effectively addresses the challenges of attribute feature representation of pixel pairs and feature extraction of attribute patterns. Experimental results show that the proposed method outperforms other approaches in terms of detection performance.
CAAI TRANSACTIONS ON INTELLIGENCE TECHNOLOGY
(2023)
Article
Geochemistry & Geophysics
Zhiyong Lv, Pingdong Zhong, Wei Wang, Zhenzhen You, Jon Atli Benediktsson, Cheng Shi
Summary: This article proposes a novel sparse key-point distance (SKPD) method based on adaptive region key-points extraction for measuring the change magnitude between bitemporal VHR_RSIs, aiming at land cover change detection (LCCD). The proposed approach includes three steps: exploring spatial-contextual information using an adaptive region generation algorithm, sparsely representing the adaptive region with the box-whisker plot theory, and defining a piecewise distance to measure the change magnitude between the bitemporal images. Experimental results based on four pairs of real VHR_RSIs and four state-of-the-art methods effectively demonstrate the superiority of the proposed approach in achieving LCCD with VHR_RSIs, with improvements in overall accuracy ranging from 5.25% to 22.24%.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Geochemistry & Geophysics
Lianru Gao, Degang Wang, Lina Zhuang, Xu Sun, Min Huang, Antonio Plaza
Summary: In this article, a new blind-spot self-supervised learning network (BS(3)LNet) is proposed for hyperspectral anomaly detection. The network is trained to reconstruct only background pixels instead of anomalous pixels by generating training patch pairs with blind spots. Experimental results show that the network is able to effectively separate anomalous pixels from the background in hyperspectral images and is competitive with other state-of-the-art approaches.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Geochemistry & Geophysics
Bing Tu, Wangquan He, Qianming Li, Yishu Peng, Antonio Plaza
Summary: Deep neural networks are important in hyperspectral image processing, but are vulnerable to adversarial samples. Existing methods do not differentiate between different classes of contextual information, leading to unreliable global context. To address this, a robust context-aware network is proposed, which generates a global contextual representation using dilated convolution and explicitly models intraclass and interclass contextual information using a class context-aware learning module. The proposed method achieves better robustness and generalization compared to other advanced techniques on benchmark HSI datasets.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Remote Sensing
Jing Ling, Shan Wei, Paolo Gamba, Rui Liu, Hongsheng Zhang
Summary: Accurate monitoring of urban impervious surfaces is crucial for understanding urbanization activities and their sustainability. Synthetic aperture radar (SAR) has the potential for urban monitoring with its all-weather imaging capability. However, identifying impervious surfaces from SAR data is challenging due to diversity and different imaging mechanisms. This study proposes a new method, polarimetric scattering mixture analysis (PSMA), to address these challenges and improve accuracy.
INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION
(2023)
Article
Geochemistry & Geophysics
Zhiyong Lv, Haitao Huang, Weiwei Sun, Tao Lei, Jon Atli Benediktsson, Junhuai Li
Summary: This paper proposes a novel approach, E-UNet, for land cover change detection with multimodal remote sensing images (MRSIs). Experimental results demonstrate the feasibility and advantages of the proposed method in terms of visual observations and quantitative evaluations.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2023)
Article
Geochemistry & Geophysics
Zhiyong Lv, Pengfei Zhang, Weiwei Sun, Jon Atli Benediktsson, Tao Lei
Summary: In this article, a novel land-cover classification method with nonparametric sample augmentation is proposed to improve the performance of HRSI classification. The method iteratively explores reliable samples and exhibits advantages in improving the visual performance and quantitative accuracies of HRSI classification.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Engineering, Electrical & Electronic
Xiaoyi Wang, Liguo Wang, Qunming Wang, Anna Vizziello, Paolo Gamba
Summary: This study proposes a method based on three-dimensional convolution and global spatial-spectral attention network to address the issue of spectral variation in hyperspectral images. A new background suppression strategy is also proposed. Experimental results show that the proposed method achieves higher accuracy in target detection.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2023)
Article
Geochemistry & Geophysics
Jia Chen, Paolo Gamba, Jun Li
Summary: In this study, we propose a multitask autoencoding model for hyperspectral unmixing. By using a 3DCNN-based network, this method enhances the robustness of the algorithm in complex environments. Additionally, it can quantitatively evaluate each area of data, improving the interpretability of the algorithm.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)