Article
Engineering, Electrical & Electronic
Muhammad Ahmad, Sidrah Shabbir, Swalpa Kumar Roy, Danfeng Hong, Xin Wu, Jing Yao, Adil Mehmood Khan, Manuel Mazzara, Salvatore Distefano, Jocelyn Chanussot
Summary: Hyperspectral imaging (HSI) is widely used in various applications, but the complex characteristics of HSI data make accurate classification challenging for traditional methods. Recent research has shown that deep learning (DL) can effectively address these challenges. This survey provides an overview of DL for HSI classification and compares state-of-the-art strategies. The article also discusses strategies to improve the generalization performance of DL in the context of limited labeled training data.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Jianlai Chen, Junchao Zhang, Yanghao Jin, Hanwen Yu, Buge Liang, De-Gui Yang
Summary: This article proposes a real-time imaging algorithm based on variable PRF for spaceborne SAR with nonlinear trajectory, which improves imaging accuracy while maintaining high efficiency.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Environmental Sciences
Rahebeh Abedi, Romulus Costache, Hossein Shafizadeh-Moghadam, Quoc Bao Pham
Summary: This article used different models to analyze the flash-flood susceptibility in the Basca Chiojdului River Basin in Romania, highlighting slope as the most important factor triggering flash floods and identifying the central part of the basin as more susceptible to flash flooding.
GEOCARTO INTERNATIONAL
(2022)
Article
Environmental Sciences
Michael A. Wulder, David P. Roy, Volker C. Radeloff, Thomas R. Loveland, Martha C. Anderson, David M. Johnson, Sean Healey, Zhe Zhu, Theodore A. Scambos, Nima Pahlevan, Matthew Hansen, Noel Gorelick, Christopher J. Crawford, Jeffrey G. Masek, Txomin Hermosilla, Joanne C. White, Alan S. Belward, Crystal Schaaf, Curtis E. Woodcock, Justin L. Huntington, Leo Lymburner, Patrick Hostert, Feng Gao, Alexei Lyapustin, Jean-Francois Pekel, Peter Strobl, Bruce D. Cook
Summary: Since 1972, the Landsat program has provided 50 years of digital, multispectral, medium spatial resolution observations, playing a crucial role in scientific and technical advancements. The program's early years brought technological breakthroughs and established a template for global Earth observation missions. The knowledge gained from Landsat has been recognized for its economic and scientific value, leading to continuous improvement and increased usage through the introduction of free and open access to data.
REMOTE SENSING OF ENVIRONMENT
(2022)
Article
Geochemistry & Geophysics
Zhaokui Li, Ming Liu, Yushi Chen, Yimin Xu, Wei Li, Qian Du
Summary: One of the challenges in hyperspectral image classification is the limited availability of labeled samples. This article introduces a novel deep cross-domain few-shot learning method (DCFSL) to address the domain adaptation and FSL issues in HSI classification. Experimental results demonstrate that DCFSL outperforms existing methods.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Mengmeng Zhang, Wei Li, Ran Tao, Hengchao Li, Qian Du
Summary: This article proposes an information fusion network called interleaving perception convolutional neural network (IP-CNN) for integrating heterogeneous information and improving joint classification performance of hyperspectral image (HSI) and light detection and ranging (LiDAR) data. The network utilizes a bidirectional autoencoder to reconstruct data and impose perception constraints for multisource structural information integration. Experimental results demonstrate the superiority of the proposed framework.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Environmental Sciences
Xiaowo Xu, Xiaoling Zhang, Tianwen Zhang
Summary: Synthetic aperture radar (SAR) satellites are widely used in maritime monitoring due to their ability to provide microwave remote sensing images without weather and light constraints. However, deploying deep learning-based SAR ship detection methods on satellites is challenging due to their complex models and high computational requirements. In this study, we propose Lite-YOLOv5, a lightweight on-board SAR ship detector based on the You Only Look Once version 5 (YOLOv5) algorithm. Lite-YOLOv5 reduces the model volume, decreases the floating-point operations (FLOPs), and achieves on-board ship detection without sacrificing accuracy.
Article
Environmental Sciences
Binh Thai Pham, Tran Van Phong, Trung Nguyen-Thoi, Kajori Parial, Sushant K. Singh, Hai-Bang Ly, Kien Trung Nguyen, Lanh Si Ho, Hiep Van Le, Indra Prakash
Summary: In this study, five spatially explicit ensemble predictive machine learning models were developed for landslide susceptibility mapping in the Van Chan district of Vietnam. The results showed that the RSSFT model performed the best in predicting future landslides and was found to be more robust than the other studied models.
GEOCARTO INTERNATIONAL
(2022)
Article
Geochemistry & Geophysics
Hongjun Su, Zhaoyue Wu, Huihui Zhang, Qian Du
IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE
(2022)
Article
Geochemistry & Geophysics
Gong Cheng, Jiabao Wang, Ke Li, Xingxing Xie, Chunbo Lang, Yanqing Yao, Junwei Han
Summary: This article proposes a novel anchor-free oriented proposal generator (AOPG) to address issues in oriented object detection caused by the use of horizontal boxes. The effectiveness of AOPG is demonstrated through extensive experiments, and a new dataset, DIOR-R, is released to alleviate the problem of data insufficiency.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Review
Environmental Sciences
Hojat Shirmard, Ehsan Farahbakhsh, R. Dietmar Muller, Rohitash Chandra
Summary: This paper comprehensively reviews the implementation and adaptation of popular and recently established machine learning methods for processing different types of remote sensing data in mineral exploration. By combining remote sensing data and machine learning methods, the capability to efficiently map critical geological features for potential maps is demonstrated. Moreover, advanced methods such as deep learning show potential to process the new generation of remote sensing data with high spatial and spectral resolution.
REMOTE SENSING OF ENVIRONMENT
(2022)
Review
Geography, Physical
Qiqi Zhu, Xi Guo, Weihuan Deng, Sunan Shi, Qingfeng Guan, Yanfei Zhong, Liangpei Zhang, Deren Li
Summary: A novel semantic change detection framework called Siam-GL was proposed for HSR remote sensing images. The Siam-GL framework effectively extracts representative features of bi-temporal HSR remote sensing images through Siamese architecture and global hierarchical sampling mechanism. Experimental results demonstrated that the Siam-GL framework outperforms advanced semantic change detection methods in terms of both quantity and quality.
ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Yifu Zhang, Peize Sun, Yi Jiang, Dongdong Yu, Fucheng Weng, Zehuan Yuan, Ping Luo, Wenyu Liu, Xinggang Wang
Summary: This method improves the performance of multi-object tracking by associating almost every detection box, effectively solving the problem of true object missing and fragmented trajectories caused by low score detection boxes being discarded. Applied to multiple trackers, this method consistently achieves improvement on IDF1 score.
COMPUTER VISION, ECCV 2022, PT XXII
(2022)
Article
Geochemistry & Geophysics
Jun Yue, Leyuan Fang, Hossein Rahmani, Pedram Ghamisi
Summary: Hyperspectral image (HSI) classification is an important topic in remote sensing, and deep learning-based methods have been widely used. However, the scarcity of labeled samples limits the potential of deep learning-based methods. To address this issue, a self-supervised learning method with adaptive distillation is proposed to train deep neural networks using extensive unlabeled samples.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Yanfeng Liu, Qiang Li, Yuan Yuan, Qian Du, Qi Wang
Summary: In this article, we propose an adaptive balanced network (ABNet) to address the challenges of remote sensing object detection. Our approach utilizes an enhanced effective channel attention mechanism (EECA), an adaptive feature pyramid network (AFPN) and a context enhancement module (CEM) to improve feature representation and capture more discriminative features. Experimental results demonstrate the superior performance of our approach.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Cheng Zhang, Wanshou Jiang, Yuan Zhang, Wei Wang, Qing Zhao, Chenjie Wang
Summary: This article presents a hybrid deep neural network that combines transformer and convolutional neural network (CNN) for semantic segmentation of very high resolution remote sensing imagery. The network utilizes a new universal backbone Swin transformer for feature extraction and incorporates various strategies for multiscale context modeling. It achieves improved accuracy through skip connections and an auxiliary boundary detection branch.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Yao Ding, Xiaofeng Zhao, Zhili Zhang, Wei Cai, Nengjun Yang
Summary: SAGE-A uses a multi-level graph sample and aggregate (graphSAGE) network to flexibly aggregate new neighbor nodes among arbitrarily structured non-Euclidean data and capture long-range contextual relations. The network utilizes the graph attention mechanism to characterize the importance among spatially neighboring regions, allowing for automatic learning of deep contextual and global information of the graph.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Geochemistry & Geophysics
Liang Zhou, Yuanxin Ye, Tengfeng Tang, Ke Nan, Yao Qin
Summary: This study employs deep learning techniques to enhance image structure features for improved matching between SAR and optical images, showing advantages over other methods in experimental results.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Geochemistry & Geophysics
Lianru Gao, Zhu Han, Danfeng Hong, Bing Zhang, Jocelyn Chanussot
Summary: In recent years, deep learning has gained attention in hyperspectral unmixing applications. However, traditional autoencoder frameworks tend to lose important detailed information. This study proposes a cycle-consistency unmixing network (CyCU-Net) to enhance the unmixing performance by learning two cascaded autoencoders and introducing a self-perception loss to achieve cycle consistency. Experimental results demonstrate that CyCU-Net outperforms other algorithms in terms of effectiveness and competitiveness.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Chunyan Yu, Rui Han, Meiping Song, Caiyu Liu, Chein-I Chang
Summary: This article presents a spatial-spectral dense CNN framework called FADCNN for hyperspectral image classification, addressing the problems of high complexity, information redundancy, and inefficient description in current networks. Experimental results show that the proposed FADCNN architecture has significant advantages compared with other state-of-the-art methods.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)