Pub Date : 2024-08-28DOI: 10.1109/JSTARS.2024.3449394
Yi Lin;Chen Gao;Jie Yu;Lang Li;Xin Chen;Yuxuan Yang;Daiqi Zhong
Global regional conflicts are on the rise, threatening both local and global stability and development. This article used monthly and daily nighttime light (NTL) imagery from December 2021 to April 2022 to analyze NTL dynamics during the initial period of the Russia–Ukraine conflict at the national and subnational (oblast and city) levels. In addition, a novel method was proposed to quantify the direct impact of the conflict at the pixel level. In this study, a new conflict effect index (CEI) was constructed to distinguish conflict-damaged area (CDA) and conflict-recovered area (CRA) and to quantify the direct damage and subsequent recovery caused by the conflict. The results indicate that the outbreak of the conflict led to a significant decrease in NTL intensity of about 56% across Ukraine. The pixel-level CEI indicated that the onset of the conflict had a broad impact across Ukraine, with the affected area shrinking as the conflict progressed. Damaged areas accounted for 12% of Ukraine at the start of the conflict, but decreased at an average rate of 2963.25 km 2 /day over the following week, while the size of recovered areas peaked at 9258.75 km 2 during the conflict, but decreased to a low of 3750 km 2 . The maximum rate of change in the area of the CDA and CRA was −67% and 135%, respectively. The study demonstrates that NTL data can capture the direct destruction of conflict and help to support humanitarian relief and postdisaster reconstruction efforts.
{"title":"Pixel-Level Quantification of Damage and Recovery Caused by the Russia–Ukraine Conflict Based on Nighttime Light Imagery","authors":"Yi Lin;Chen Gao;Jie Yu;Lang Li;Xin Chen;Yuxuan Yang;Daiqi Zhong","doi":"10.1109/JSTARS.2024.3449394","DOIUrl":"https://doi.org/10.1109/JSTARS.2024.3449394","url":null,"abstract":"Global regional conflicts are on the rise, threatening both local and global stability and development. This article used monthly and daily nighttime light (NTL) imagery from December 2021 to April 2022 to analyze NTL dynamics during the initial period of the Russia–Ukraine conflict at the national and subnational (oblast and city) levels. In addition, a novel method was proposed to quantify the direct impact of the conflict at the pixel level. In this study, a new conflict effect index (CEI) was constructed to distinguish conflict-damaged area (CDA) and conflict-recovered area (CRA) and to quantify the direct damage and subsequent recovery caused by the conflict. The results indicate that the outbreak of the conflict led to a significant decrease in NTL intensity of about 56% across Ukraine. The pixel-level CEI indicated that the onset of the conflict had a broad impact across Ukraine, with the affected area shrinking as the conflict progressed. Damaged areas accounted for 12% of Ukraine at the start of the conflict, but decreased at an average rate of 2963.25 km\u0000<sup>2</sup>\u0000/day over the following week, while the size of recovered areas peaked at 9258.75 km\u0000<sup>2</sup>\u0000 during the conflict, but decreased to a low of 3750 km\u0000<sup>2</sup>\u0000. The maximum rate of change in the area of the CDA and CRA was −67% and 135%, respectively. The study demonstrates that NTL data can capture the direct destruction of conflict and help to support humanitarian relief and postdisaster reconstruction efforts.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":null,"pages":null},"PeriodicalIF":4.7,"publicationDate":"2024-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10654582","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142143790","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Optimal use of agricultural land leads to increased productivity and paves the way for sustainable agriculture. The land suitability assessment (LSA) is known as the basic scientific solution in this regard and has been widely used. This study takes a new approach to LSA based on the cultivated crops and their land production potential (LPP). The Honam region of Lorestan Province in western Iran was selected as a case study, where various plants including orchards and crop types (e.g., rainfed wheat, irrigated wheat, barley, chickpea, alfalfa, clover, and garlic) were cultivated. A map of crop types was first produced using the time series of Sentinel-1 synthetic aperture radar and Sentinel-2 optical imagery in the context of a feature-level fusion classification approach based on the support vector machine (SVM) algorithm. The LSA was then performed using the Food and Agriculture Organization method, and the LPP for plants was estimated using climate, soil, and landscape requirements. The SVM-derived crop type map gave acceptable performance with an overall accuracy of 92% and 86% kappa coefficient. Meanwhile, the clover and alfalfa had the highest (96%) and lowest (66%) accuracy, respectively. The LSA showed that slope, temperature, and soil physical properties such as texture, structure, and coarse fragments resulted in a significant reduction in LPP compared to potential yield. These limitations lowered the land suitability for chickpea, wheat, and garlic and made the land unsuitable for chickpea. Despite the limitations, over 60% of the studied lands for alfalfa, clover, barley, and wheat were in the suitable and moderately suitable classes.
{"title":"Land Suitability Assessment Based on Feature-Level Fusion of Sentinel-1 and Sentinel-2 Imagery: A Case Study of the Honam Region of Iran","authors":"Bahare Delsous Khaki;Mansour Chatrenour;Mir Naser Navidi;Masoud Soleimani;Saham Mirzaei;Stefano Pignatti","doi":"10.1109/JSTARS.2024.3437689","DOIUrl":"https://doi.org/10.1109/JSTARS.2024.3437689","url":null,"abstract":"Optimal use of agricultural land leads to increased productivity and paves the way for sustainable agriculture. The land suitability assessment (LSA) is known as the basic scientific solution in this regard and has been widely used. This study takes a new approach to LSA based on the cultivated crops and their land production potential (LPP). The Honam region of Lorestan Province in western Iran was selected as a case study, where various plants including orchards and crop types (e.g., rainfed wheat, irrigated wheat, barley, chickpea, alfalfa, clover, and garlic) were cultivated. A map of crop types was first produced using the time series of Sentinel-1 synthetic aperture radar and Sentinel-2 optical imagery in the context of a feature-level fusion classification approach based on the support vector machine (SVM) algorithm. The LSA was then performed using the Food and Agriculture Organization method, and the LPP for plants was estimated using climate, soil, and landscape requirements. The SVM-derived crop type map gave acceptable performance with an overall accuracy of 92% and 86% kappa coefficient. Meanwhile, the clover and alfalfa had the highest (96%) and lowest (66%) accuracy, respectively. The LSA showed that slope, temperature, and soil physical properties such as texture, structure, and coarse fragments resulted in a significant reduction in LPP compared to potential yield. These limitations lowered the land suitability for chickpea, wheat, and garlic and made the land unsuitable for chickpea. Despite the limitations, over 60% of the studied lands for alfalfa, clover, barley, and wheat were in the suitable and moderately suitable classes.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":null,"pages":null},"PeriodicalIF":4.7,"publicationDate":"2024-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10653750","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142143710","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-26DOI: 10.1109/JSTARS.2024.3449441
Tiantian Tang;Tao Chen;Guan Gui
In machine learning (ML)-based hydrological forecasting, particularly in medium- and long-term prediction, judicious predictor selection is paramount, as it ultimately determines the forecast accuracy. This study pioneered an advanced predictor-screening method that synergizes the mutual information (MI) and random forest (RF) technologies through minimum-redundancy-maximum-relevance-recursive feature elimination-random forest (mRMR-RFE-RF) method, blending both filtering and wrapping techniques. This method was rigorously tested through a detailed case study in the Danjiangkou basin, where a comprehensive analysis of 1560 meteorological factors was conducted. Employing three sophisticated ML algorithms—RF, eXtreme Gradient Boosting (XGB), and Light Gradient Boosting (LGB)—we developed precipitation forecasting models. Furthermore, we performed an in-depth rationality analysis of high-frequency predictors. The findings from our study show that this novel hybrid screening strategy markedly outperformed conventional singular predictor-screening methods in enhancing the accuracy of precipitation forecasting when integrated into these forecasting models. Moreover, it assured the validity of the high-frequency forecast factors employed. Therefore, this innovative method not only elevates the accuracy of medium- and long-term precipitation forecasting but also contributes a novel perspective to the methodology of predictor selection in hydrological forecasting models.
{"title":"A Novel mRMR-RFE-RF Method for Enhancing Medium- and Long-Term Hydrological Forecasting: A Case Study of the Danjiangkou Basin","authors":"Tiantian Tang;Tao Chen;Guan Gui","doi":"10.1109/JSTARS.2024.3449441","DOIUrl":"https://doi.org/10.1109/JSTARS.2024.3449441","url":null,"abstract":"In machine learning (ML)-based hydrological forecasting, particularly in medium- and long-term prediction, judicious predictor selection is paramount, as it ultimately determines the forecast accuracy. This study pioneered an advanced predictor-screening method that synergizes the mutual information (MI) and random forest (RF) technologies through minimum-redundancy-maximum-relevance-recursive feature elimination-random forest (mRMR-RFE-RF) method, blending both filtering and wrapping techniques. This method was rigorously tested through a detailed case study in the Danjiangkou basin, where a comprehensive analysis of 1560 meteorological factors was conducted. Employing three sophisticated ML algorithms—RF, eXtreme Gradient Boosting (XGB), and Light Gradient Boosting (LGB)—we developed precipitation forecasting models. Furthermore, we performed an in-depth rationality analysis of high-frequency predictors. The findings from our study show that this novel hybrid screening strategy markedly outperformed conventional singular predictor-screening methods in enhancing the accuracy of precipitation forecasting when integrated into these forecasting models. Moreover, it assured the validity of the high-frequency forecast factors employed. Therefore, this innovative method not only elevates the accuracy of medium- and long-term precipitation forecasting but also contributes a novel perspective to the methodology of predictor selection in hydrological forecasting models.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":null,"pages":null},"PeriodicalIF":4.7,"publicationDate":"2024-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10646493","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142143713","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-26DOI: 10.1109/JSTARS.2024.3449436
Hanhu Liu;Heng Zhang;Ronghao Yang
Although many scholars have realized that deep learning methods have great advantages in hyperspectral lithology classification now, most of them use simple convolutional neural networks for discussion, which are difficult to effectively extract spectral sequence information from hyperspectral image (HSI). To address this issue, we propose a double-branch multiscale dual-attention (DBMSDA) network using two HSIs multiple different rock types as data sources. This network utilizes the multiscale spectral residual self-attention structure and dense connections in the spectral branch to extract diagnostic multiscale spectral information from HSIs, improving the ability to reuse spectral information. In the spatial branch, dense connections are used to fully extract diagnostic spatial information, and finally, the two are fused for classification. We compare the DBMSDA network with several traditional machine learning and deep learning methods. Experimental results show that the DBMSDA network has the best lithology classification performance and robustness. Its OA on the ZCD1 and HCD2 datasets is 85.08% and 90.80%, respectively, representing an improvement of 5.95% and 2.4% over the lowest OA, confirming that the DBMSDA network can effectively improve accuracy and is superior to other methods. Meanwhile, the experiment also showed that the DBMSDA network has good classification performance on HSIs of various rock types and HSIs from different data sources. Finally, ablation experiments were conducted to demonstrate that the DBMSDA network relies more on large training samples compared with traditional machine learning models, and that the DBMSDA network has better applicability and stability under the same sufficient training samples.
{"title":"Lithological Classification by Hyperspectral Remote Sensing Images Based on Double-Branch Multiscale Dual-Attention Network","authors":"Hanhu Liu;Heng Zhang;Ronghao Yang","doi":"10.1109/JSTARS.2024.3449436","DOIUrl":"https://doi.org/10.1109/JSTARS.2024.3449436","url":null,"abstract":"Although many scholars have realized that deep learning methods have great advantages in hyperspectral lithology classification now, most of them use simple convolutional neural networks for discussion, which are difficult to effectively extract spectral sequence information from hyperspectral image (HSI). To address this issue, we propose a double-branch multiscale dual-attention (DBMSDA) network using two HSIs multiple different rock types as data sources. This network utilizes the multiscale spectral residual self-attention structure and dense connections in the spectral branch to extract diagnostic multiscale spectral information from HSIs, improving the ability to reuse spectral information. In the spatial branch, dense connections are used to fully extract diagnostic spatial information, and finally, the two are fused for classification. We compare the DBMSDA network with several traditional machine learning and deep learning methods. Experimental results show that the DBMSDA network has the best lithology classification performance and robustness. Its OA on the ZCD1 and HCD2 datasets is 85.08% and 90.80%, respectively, representing an improvement of 5.95% and 2.4% over the lowest OA, confirming that the DBMSDA network can effectively improve accuracy and is superior to other methods. Meanwhile, the experiment also showed that the DBMSDA network has good classification performance on HSIs of various rock types and HSIs from different data sources. Finally, ablation experiments were conducted to demonstrate that the DBMSDA network relies more on large training samples compared with traditional machine learning models, and that the DBMSDA network has better applicability and stability under the same sufficient training samples.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":null,"pages":null},"PeriodicalIF":4.7,"publicationDate":"2024-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10646515","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142143762","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
To overcome the effect of grid mismatch on the superresolution performance, a sparse Bayesian learning-based multichannel radar forward-looking superresolution imaging scheme is proposed in this article. In the scheme, a coarse imaging grid is initialized first, and local grid refinement is performed based on the preliminary estimation results of the target information. Then, based on the refined grid, an off-grid superresolution model considering grid mismatch is established, and the total least squares method is used to estimate the mismatch error for modifying the steering matrix in superresolution processing. At last, based on the modified steering matrix, sparse Bayesian learning algorithm is iteratively executed to achieve multichannel radar forward-looking superresolution imaging. Simulated and measured data processing results are illustrated to verify the effectiveness of the proposed scheme.
{"title":"Sparse Bayesian Learning-Based Multichannel Radar Forward-Looking Superresolution Imaging Considering Grid Mismatch","authors":"Jianyu Yang;Wenchao Li;Kefeng Li;Rui Chen;Kun Zhang;Deqing Mao;Yin Zhang","doi":"10.1109/JSTARS.2024.3448365","DOIUrl":"https://doi.org/10.1109/JSTARS.2024.3448365","url":null,"abstract":"To overcome the effect of grid mismatch on the superresolution performance, a sparse Bayesian learning-based multichannel radar forward-looking superresolution imaging scheme is proposed in this article. In the scheme, a coarse imaging grid is initialized first, and local grid refinement is performed based on the preliminary estimation results of the target information. Then, based on the refined grid, an off-grid superresolution model considering grid mismatch is established, and the total least squares method is used to estimate the mismatch error for modifying the steering matrix in superresolution processing. At last, based on the modified steering matrix, sparse Bayesian learning algorithm is iteratively executed to achieve multichannel radar forward-looking superresolution imaging. Simulated and measured data processing results are illustrated to verify the effectiveness of the proposed scheme.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":null,"pages":null},"PeriodicalIF":4.7,"publicationDate":"2024-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10645205","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142143606","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-23DOI: 10.1109/JSTARS.2024.3448355
Yuting Qi;Lei Zhong;Yaoming Ma;Yunfei Fu;Zixin Wang;Peizhen Li
Land surface temperature (LST) is a crucial physical parameter for hydrological, meteorological, climatological, and climate change studies. To encourage the use of satellite-derived LST products in a wide range of applications, providing feedback on product performance over regional and global scales is an urgent task. However, considering that the uncertainty of newly released LST products is still unclear, it is urgently necessary to perform a comprehensive validation and error analysis, especially in areas with special geographical and weather conditions, such as the Tibetan plateau (TP). In particular, fewer studies have been concerned with the degraded LST retrieval accuracy over the TP because of the sparse ground measurements. In this study, moderate-resolution imaging spectroradiometer (MODIS) LST products (C6.1) were comprehensively evaluated based on the independent ground observation systems with different atmospheric and LST conditions. The in situ measurements collected from the Tibetan Observation and Research Platform and surface radiation systems are located on the American Plain and the TP, respectively, incorporating various land-cover types, including barren land, grassland, cropland, shrubland, and sparse and dense vegetation, among others. The spatial representativeness evaluation indicated that relatively high-quality in situ LSTs can be obtained during nighttime. Compared with the North American Plain (with a mean RMSE of 1.56 K), MODIS LST retrievals have larger discrepancies (mean RMSE of 2.34 K) over the TP with complex terrain and weather conditions. Emissivity determination is the primary source of the uncertainty in the generalized split-window (GSW) algorithm. Moreover, simulation settings of atmospheric and LST conditions in the GSW algorithm cannot cover a wide range of conditions at a global scale. It is expected to develop new LST retrieval algorithm to meet the quality specifications of users over the TP. Overall, this study identifies critical further research needs and improves the understanding of LST product performance under complex circumstances.
{"title":"Evaluation of MODIS LST Products Over the Tibetan Plateau and Plain Areas With in Situ Measurements","authors":"Yuting Qi;Lei Zhong;Yaoming Ma;Yunfei Fu;Zixin Wang;Peizhen Li","doi":"10.1109/JSTARS.2024.3448355","DOIUrl":"https://doi.org/10.1109/JSTARS.2024.3448355","url":null,"abstract":"Land surface temperature (LST) is a crucial physical parameter for hydrological, meteorological, climatological, and climate change studies. To encourage the use of satellite-derived LST products in a wide range of applications, providing feedback on product performance over regional and global scales is an urgent task. However, considering that the uncertainty of newly released LST products is still unclear, it is urgently necessary to perform a comprehensive validation and error analysis, especially in areas with special geographical and weather conditions, such as the Tibetan plateau (TP). In particular, fewer studies have been concerned with the degraded LST retrieval accuracy over the TP because of the sparse ground measurements. In this study, moderate-resolution imaging spectroradiometer (MODIS) LST products (C6.1) were comprehensively evaluated based on the independent ground observation systems with different atmospheric and LST conditions. The in situ measurements collected from the Tibetan Observation and Research Platform and surface radiation systems are located on the American Plain and the TP, respectively, incorporating various land-cover types, including barren land, grassland, cropland, shrubland, and sparse and dense vegetation, among others. The spatial representativeness evaluation indicated that relatively high-quality in situ LSTs can be obtained during nighttime. Compared with the North American Plain (with a mean RMSE of 1.56 K), MODIS LST retrievals have larger discrepancies (mean RMSE of 2.34 K) over the TP with complex terrain and weather conditions. Emissivity determination is the primary source of the uncertainty in the generalized split-window (GSW) algorithm. Moreover, simulation settings of atmospheric and LST conditions in the GSW algorithm cannot cover a wide range of conditions at a global scale. It is expected to develop new LST retrieval algorithm to meet the quality specifications of users over the TP. Overall, this study identifies critical further research needs and improves the understanding of LST product performance under complex circumstances.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":null,"pages":null},"PeriodicalIF":4.7,"publicationDate":"2024-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10645202","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142143711","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-23DOI: 10.1109/JSTARS.2024.3448430
Hongkang Zhang;Shao-Lun Huang;Ercan Engin Kuruoglu
In the realm of urban development, the precise classification and identification of land types are crucial for improving land use efficiency. This article proposes a land recognition and classification method based on data sparsity and improved Soft Hirschfeld-Gebelein-Rényi (Soft-HGR) under multimodal conditions. First, a sparse information processing module is designed to enhance information accuracy and quickly obtain data sample features. Then, to solve the problem of information independence in single mode and lack of fusion in multimodal mode, an improved SoftHGR module is developed. This module incorporates covariance and trace constraints, enhances machine learning efficiency by stabilizing output and addressing dimensionality and variance issues in HGR, and speeds up land classification by cross-fusing multimodal features to deepen the understanding of diverse information interconnections. Based on this, a multimodal MI-SoftHGR fusion network is constructed, which can achieve cross-correlation sharing and collaborative extraction of feature information, thereby realizing accurate remote sensing image recognition and classification under multimodal conditions. Finally, empirical evaluations were conducted on Berlin, Augsburg, and MUUFL datasets, and the proposed method was compared with state-of-the-art algorithms. The results fully validate the efficacy and significant superiority of the proposed method.
{"title":"MHFNet: An Improved HGR Multimodal Network for Informative Correlation Fusion in Remote Sensing Image Classification","authors":"Hongkang Zhang;Shao-Lun Huang;Ercan Engin Kuruoglu","doi":"10.1109/JSTARS.2024.3448430","DOIUrl":"https://doi.org/10.1109/JSTARS.2024.3448430","url":null,"abstract":"In the realm of urban development, the precise classification and identification of land types are crucial for improving land use efficiency. This article proposes a land recognition and classification method based on data sparsity and improved Soft Hirschfeld-Gebelein-Rényi (Soft-HGR) under multimodal conditions. First, a sparse information processing module is designed to enhance information accuracy and quickly obtain data sample features. Then, to solve the problem of information independence in single mode and lack of fusion in multimodal mode, an improved SoftHGR module is developed. This module incorporates covariance and trace constraints, enhances machine learning efficiency by stabilizing output and addressing dimensionality and variance issues in HGR, and speeds up land classification by cross-fusing multimodal features to deepen the understanding of diverse information interconnections. Based on this, a multimodal MI-SoftHGR fusion network is constructed, which can achieve cross-correlation sharing and collaborative extraction of feature information, thereby realizing accurate remote sensing image recognition and classification under multimodal conditions. Finally, empirical evaluations were conducted on Berlin, Augsburg, and MUUFL datasets, and the proposed method was compared with state-of-the-art algorithms. The results fully validate the efficacy and significant superiority of the proposed method.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":null,"pages":null},"PeriodicalIF":4.7,"publicationDate":"2024-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10645073","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142143641","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-22DOI: 10.1109/JSTARS.2024.3447896
Linwei Li;Ziyu Wu;Bin Wang
Anomaly detection (AD) aiming to locate targets distinct from the surrounding background spectra remains a challenging task in hyperspectral applications. The methods based on low-rank decomposition utilize the inherent low-rank characteristic of hyperspectral images (HSIs), which has attracted great interest and achieved many advances in recent years. In order to fully consider the characteristics of HSIs, more appropriate constrains need to be added to the low-rank model. However, there are too many regularizations and mutual constraints between regularizers, which would result in a reduction in detection accuracy, while an increasing number of tradeoff parameters complicates parameter tuning. To address the above problems, we propose a novel method based on merging total variation into low-rank representation (MTVLRR) for hyperspectral AD in this article, using a regularizer to reflect the low-rankness and smoothness of the background component of HSIs simultaneously, which can significantly decrease the mutual influence of regularizers and the difficulty of parameter tuning. Experimental results on both simulated and real hyperspectral datasets demonstrate that the proposed MTVLRR has an excellent AD performance in terms of detection accuracy compared with other state-of-the-art methods.
异常检测（AD）旨在定位与周围背景光谱不同的目标，这在高光谱应用中仍然是一项具有挑战性的任务。基于低秩分解的方法利用了高光谱图像（HSI）固有的低秩特征，近年来引起了人们的极大兴趣，并取得了许多进展。为了充分考虑高光谱图像的特性，需要在低秩模型中加入更多适当的约束条件。然而，正则化和正则化之间的相互约束过多，会导致检测精度下降，而权衡参数的增加又使参数调整变得复杂。针对上述问题，我们在本文中提出了一种基于将总变异合并为低秩表示（MTVLRR）的高光谱 AD 新方法，利用正则化器同时反映 HSI 的低秩性和背景成分的平滑性，可以显著降低正则化器之间的相互影响和参数调优的难度。在模拟和真实高光谱数据集上的实验结果表明，与其他最先进的方法相比，所提出的 MTVLRR 在检测精度方面具有出色的 AD 性能。
{"title":"Hyperspectral Anomaly Detection via Merging Total Variation Into Low-Rank Representation","authors":"Linwei Li;Ziyu Wu;Bin Wang","doi":"10.1109/JSTARS.2024.3447896","DOIUrl":"https://doi.org/10.1109/JSTARS.2024.3447896","url":null,"abstract":"Anomaly detection (AD) aiming to locate targets distinct from the surrounding background spectra remains a challenging task in hyperspectral applications. The methods based on low-rank decomposition utilize the inherent low-rank characteristic of hyperspectral images (HSIs), which has attracted great interest and achieved many advances in recent years. In order to fully consider the characteristics of HSIs, more appropriate constrains need to be added to the low-rank model. However, there are too many regularizations and mutual constraints between regularizers, which would result in a reduction in detection accuracy, while an increasing number of tradeoff parameters complicates parameter tuning. To address the above problems, we propose a novel method based on merging total variation into low-rank representation (MTVLRR) for hyperspectral AD in this article, using a regularizer to reflect the low-rankness and smoothness of the background component of HSIs simultaneously, which can significantly decrease the mutual influence of regularizers and the difficulty of parameter tuning. Experimental results on both simulated and real hyperspectral datasets demonstrate that the proposed MTVLRR has an excellent AD performance in terms of detection accuracy compared with other state-of-the-art methods.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":null,"pages":null},"PeriodicalIF":4.7,"publicationDate":"2024-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10643646","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142143589","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-21DOI: 10.1109/JSTARS.2024.3446815
Mingyao Chen;Tianpeng Liu;Li Liu
In recent years, synthetic aperture radar automatic target recognition (SAR ATR) has been widely researched for its ability to achieve high-performance target classification through supervised training, facilitating battlefield reconnaissance, and intelligence generation. When dealing with unknown class data for ATR model training, significant time and effort are typically required for manual interpretation and labeling. However, when unknown class data shares the same domain as the known labeled data in the library, leveraging their shared deep semantic knowledge can enable automatic classification and labeling of the unknown class data. In this article, we investigate the novel category discovery (NCD) problem in SAR images, using labeled data to guide the clustering process of new class data. Specifically, we utilize the Unified Objective function training framework to address the training imbalance between labeled and unlabeled data in the NCD process, incorporating various improvements on this foundation. Through a binary segmentation-based strategy, we effectively mitigate the interference of “noise pairs” with significant semantic differences on the model's self-supervised pretraining. In addition, we introduce multicrop consistency loss and equal distance loss to impose constraints on training by leveraging intraclass and interclass relationships in the latent space, thereby obtaining representations with higher interclass separability. Our method achieves state-of-the-art clustering performance in multiple scenarios. Extensive experimental results on the MSTAR benchmark dataset demonstrate the effectiveness of the proposed methods.
{"title":"A Novel Category Discovery Method for SAR Images Based on an Improved UNO Framework","authors":"Mingyao Chen;Tianpeng Liu;Li Liu","doi":"10.1109/JSTARS.2024.3446815","DOIUrl":"https://doi.org/10.1109/JSTARS.2024.3446815","url":null,"abstract":"In recent years, synthetic aperture radar automatic target recognition (SAR ATR) has been widely researched for its ability to achieve high-performance target classification through supervised training, facilitating battlefield reconnaissance, and intelligence generation. When dealing with unknown class data for ATR model training, significant time and effort are typically required for manual interpretation and labeling. However, when unknown class data shares the same domain as the known labeled data in the library, leveraging their shared deep semantic knowledge can enable automatic classification and labeling of the unknown class data. In this article, we investigate the novel category discovery (NCD) problem in SAR images, using labeled data to guide the clustering process of new class data. Specifically, we utilize the Unified Objective function training framework to address the training imbalance between labeled and unlabeled data in the NCD process, incorporating various improvements on this foundation. Through a binary segmentation-based strategy, we effectively mitigate the interference of “noise pairs” with significant semantic differences on the model's self-supervised pretraining. In addition, we introduce multicrop consistency loss and equal distance loss to impose constraints on training by leveraging intraclass and interclass relationships in the latent space, thereby obtaining representations with higher interclass separability. Our method achieves state-of-the-art clustering performance in multiple scenarios. Extensive experimental results on the MSTAR benchmark dataset demonstrate the effectiveness of the proposed methods.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":null,"pages":null},"PeriodicalIF":4.7,"publicationDate":"2024-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10643254","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142143708","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Evapotranspiration (ET) is a crucial parameter in agriculture as it plays a vital role in managing water resources, monitoring droughts, and optimizing crop yields across different ecosystems. Given its significance in crop growth, it is essential to measure ET accurately and continuously to conduct precise analyses in agriculture. However, the continuous monitoring of ET changes is very challenging: while in-situ measurements are costly and not feasible for covering a wide geography, remote sensing-based ET products are typically dependent on optical satellites that cannot operate and transmit data under certain weather conditions, especially in the presence of clouds. In this article, we present the first comprehensive study on predicting ET from synthetic aperture radar (SAR) imagery, which we refer to as SAR2ET. Our work is motivated by the fact that SAR has the critical advantages of being all-weather available and sensitive to crop and soil changes. In handling the SAR2ET problem, we additionally incorporate nonoptical meteorological and topographical input data from auxiliary data sources. We approach SAR2ET as a multimodal image-to-image translation task, for which we train a UNet-shaped network. To evaluate the effectiveness of SAR-based ET predictions, we construct a benchmark dataset over a large geographical region with image samples covering a whole agriculture season. Our experimental findings on this dataset suggest that first, the proposed approach leads to strong results, second, valuable information can be extracted from both SAR and auxiliary data sources, and finally, SAR2ET is overall a promising research direction toward obtaining data-driven year-round ET estimates. The benchmark dataset will be shared publicly upon publication to stimulate future work.
{"title":"SAR2ET: End-to-End SAR-Driven Multisource ET Imagery Estimation Over Croplands","authors":"Samet Cetin;Berk Ülker;Esra Erten;Ramazan Gokberk Cinbis","doi":"10.1109/JSTARS.2024.3447033","DOIUrl":"https://doi.org/10.1109/JSTARS.2024.3447033","url":null,"abstract":"Evapotranspiration (ET) is a crucial parameter in agriculture as it plays a vital role in managing water resources, monitoring droughts, and optimizing crop yields across different ecosystems. Given its significance in crop growth, it is essential to measure ET accurately and continuously to conduct precise analyses in agriculture. However, the continuous monitoring of ET changes is very challenging: while in-situ measurements are costly and not feasible for covering a wide geography, remote sensing-based ET products are typically dependent on optical satellites that cannot operate and transmit data under certain weather conditions, especially in the presence of clouds. In this article, we present the first comprehensive study on predicting ET from synthetic aperture radar (SAR) imagery, which we refer to as SAR2ET. Our work is motivated by the fact that SAR has the critical advantages of being all-weather available and sensitive to crop and soil changes. In handling the SAR2ET problem, we additionally incorporate nonoptical meteorological and topographical input data from auxiliary data sources. We approach SAR2ET as a multimodal image-to-image translation task, for which we train a UNet-shaped network. To evaluate the effectiveness of SAR-based ET predictions, we construct a benchmark dataset over a large geographical region with image samples covering a whole agriculture season. Our experimental findings on this dataset suggest that first, the proposed approach leads to strong results, second, valuable information can be extracted from both SAR and auxiliary data sources, and finally, SAR2ET is overall a promising research direction toward obtaining data-driven year-round ET estimates. The benchmark dataset will be shared publicly upon publication to stimulate future work.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":null,"pages":null},"PeriodicalIF":4.7,"publicationDate":"2024-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10643277","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142143590","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：