Pub Date : 2024-11-25DOI: 10.1109/LGRS.2024.3505294
Ziyi Wang;Feng Gao;Junyu Dong;Qian Du
Recently, transformer-based hyperspectral image (HSI) change detection methods have shown remarkable performance. Nevertheless, existing attention mechanisms in transformers have limitations in local feature representation. To address this issue, we propose global and local attention-based transformer (GLAFormer), which incorporates a global and local attention module (GLAM) to combine high-frequency and low-frequency signals. Furthermore, we introduce a cross-gating mechanism, called cross-gated feedforward network (CGFN), to emphasize salient features and suppress noise interference. Specifically, the GLAM splits attention heads into global and local attention components to capture comprehensive spatial–spectral features. The global attention component uses global attention on downsampled feature maps to capture low-frequency information, while the local attention component focuses on high-frequency details using nonoverlapping window-based local attention. The CGFN enhances the feature representation via convolutions and cross-gating mechanism in parallel paths. The proposed GLAFormer is evaluated on three HSI datasets. The results demonstrate its superiority over state-of-the-art HSI change detection methods. The source code of GLAFormer is available at https://github.com/summitgao/GLAFormer .
{"title":"Global and Local Attention-Based Transformer for Hyperspectral Image Change Detection","authors":"Ziyi Wang;Feng Gao;Junyu Dong;Qian Du","doi":"10.1109/LGRS.2024.3505294","DOIUrl":"https://doi.org/10.1109/LGRS.2024.3505294","url":null,"abstract":"Recently, transformer-based hyperspectral image (HSI) change detection methods have shown remarkable performance. Nevertheless, existing attention mechanisms in transformers have limitations in local feature representation. To address this issue, we propose global and local attention-based transformer (GLAFormer), which incorporates a global and local attention module (GLAM) to combine high-frequency and low-frequency signals. Furthermore, we introduce a cross-gating mechanism, called cross-gated feedforward network (CGFN), to emphasize salient features and suppress noise interference. Specifically, the GLAM splits attention heads into global and local attention components to capture comprehensive spatial–spectral features. The global attention component uses global attention on downsampled feature maps to capture low-frequency information, while the local attention component focuses on high-frequency details using nonoverlapping window-based local attention. The CGFN enhances the feature representation via convolutions and cross-gating mechanism in parallel paths. The proposed GLAFormer is evaluated on three HSI datasets. The results demonstrate its superiority over state-of-the-art HSI change detection methods. The source code of GLAFormer is available at \u0000<uri>https://github.com/summitgao/GLAFormer</uri>\u0000.","PeriodicalId":91017,"journal":{"name":"IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society","volume":"22 ","pages":"1-5"},"PeriodicalIF":0.0,"publicationDate":"2024-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142810455","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The passive surface wave dispersion imaging is extensively utilized for shallow surface velocity inversion. However, the presence of strong directional noise sources often leads to deviations from the truth dispersion. Conventional beamforming technique can correct dispersion spectrum, but with limited resolution. Additionally, actual records contain random noise, which further compromises imaging quality. To address these challenges concerning dispersion imaging resolution and noise resistance, we propose a high-resolution dispersion imaging method that integrates the multiple signal classification (MUSIC) algorithm with subarray spatial smoothing processing. Initially, velocity is incorporated into the MUSIC algorithm to discern the direction of ambient noise, thereby extracting a sparse f–v spectrum free from random noise interference. To further mitigate the impact of random noise, a subarray spatial-smoothing MUSIC approach is devised, effectively reducing such interferences. Synthetic and field experiments demonstrate its capability to achieve high-resolution dispersion spectrum even in the presence of noise.
{"title":"High-Resolution Directional Passive Surface Waves Dispersion Imaging Based on Smoothing MUSIC","authors":"Yaru Xue;Qi Liang;Jingjie Cao;Ming Jiang;Luyu Feng;Junli Su;Cheng Zhang","doi":"10.1109/LGRS.2024.3506165","DOIUrl":"https://doi.org/10.1109/LGRS.2024.3506165","url":null,"abstract":"The passive surface wave dispersion imaging is extensively utilized for shallow surface velocity inversion. However, the presence of strong directional noise sources often leads to deviations from the truth dispersion. Conventional beamforming technique can correct dispersion spectrum, but with limited resolution. Additionally, actual records contain random noise, which further compromises imaging quality. To address these challenges concerning dispersion imaging resolution and noise resistance, we propose a high-resolution dispersion imaging method that integrates the multiple signal classification (MUSIC) algorithm with subarray spatial smoothing processing. Initially, velocity is incorporated into the MUSIC algorithm to discern the direction of ambient noise, thereby extracting a sparse f–v spectrum free from random noise interference. To further mitigate the impact of random noise, a subarray spatial-smoothing MUSIC approach is devised, effectively reducing such interferences. Synthetic and field experiments demonstrate its capability to achieve high-resolution dispersion spectrum even in the presence of noise.","PeriodicalId":91017,"journal":{"name":"IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society","volume":"22 ","pages":"1-5"},"PeriodicalIF":0.0,"publicationDate":"2024-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142789104","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-22DOI: 10.1109/LGRS.2024.3505139
Asem Melebari;Sepehr Eskandari;Mahta Moghaddam
Recent advancements in the uncrewed aerial vehicles (UAVs) as remote system platforms have enabled low-cost and easy-to-use options for various applications. An example of such a remote sensing system is a software-defined radar (SDRadar) mounted on a UAV. There is a need to accurately calibrate the radar system with low-cost and field-ready schemes to utilize the full potential of the radar system. Our system is an SDRadar mounted on a small UAV. The waveform of the radar has a 1-GHz bandwidth with a center frequency set to 750 MHz. The calibration employs a 59-cm-diameter disk elevated above the ground as an external passive target. Our results demonstrate a calibration accuracy of approximately 1 dB across most of the transmitted band, in both laboratory and field conditions. However, calibration challenges persist in the lowest parts of the frequency range, which are affected by the high side lobes and grating lobes in the antenna pattern. This calibration process is adaptable to various bandwidths and center frequencies.
{"title":"Absolute RCS Calibration of a UAV Ultrawideband Surface Penetrating Radar Using a Disk","authors":"Asem Melebari;Sepehr Eskandari;Mahta Moghaddam","doi":"10.1109/LGRS.2024.3505139","DOIUrl":"https://doi.org/10.1109/LGRS.2024.3505139","url":null,"abstract":"Recent advancements in the uncrewed aerial vehicles (UAVs) as remote system platforms have enabled low-cost and easy-to-use options for various applications. An example of such a remote sensing system is a software-defined radar (SDRadar) mounted on a UAV. There is a need to accurately calibrate the radar system with low-cost and field-ready schemes to utilize the full potential of the radar system. Our system is an SDRadar mounted on a small UAV. The waveform of the radar has a 1-GHz bandwidth with a center frequency set to 750 MHz. The calibration employs a 59-cm-diameter disk elevated above the ground as an external passive target. Our results demonstrate a calibration accuracy of approximately 1 dB across most of the transmitted band, in both laboratory and field conditions. However, calibration challenges persist in the lowest parts of the frequency range, which are affected by the high side lobes and grating lobes in the antenna pattern. This calibration process is adaptable to various bandwidths and center frequencies.","PeriodicalId":91017,"journal":{"name":"IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society","volume":"22 ","pages":"1-5"},"PeriodicalIF":0.0,"publicationDate":"2024-11-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142821188","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-20DOI: 10.1109/LGRS.2024.3504293
Nikolaos Dionelis;Nicolas Longépé
Confidence assessments of semantic segmentation algorithms are important. Ideally, models should have the ability to predict in advance whether their output is likely to be incorrect. Assessing the confidence levels of model predictions in Earth observation (EO) classification is essential, as it can enhance semantic segmentation performance and help prevent further exploitation of the results in the case of erroneous prediction. The model we developed, Confidence Assessment for enhanced Semantic segmentation (CAS), evaluates confidence at both the segment and pixel levels, providing both labels and confidence scores as output. Our model, CAS, identifies segments with incorrectly predicted labels using the proposed combined confidence metric, refines the model, and enhances its performance. This work has significant applications, particularly in evaluating EO Foundation Models on semantic segmentation downstream tasks, such as land-cover classification using Sentinel-2 satellite data. The evaluation results show that this strategy is effective and that the proposed model CAS outperforms other baseline models.
{"title":"Fine-Tuning Foundation Models With Confidence Assessment for Enhanced Semantic Segmentation","authors":"Nikolaos Dionelis;Nicolas Longépé","doi":"10.1109/LGRS.2024.3504293","DOIUrl":"https://doi.org/10.1109/LGRS.2024.3504293","url":null,"abstract":"Confidence assessments of semantic segmentation algorithms are important. Ideally, models should have the ability to predict in advance whether their output is likely to be incorrect. Assessing the confidence levels of model predictions in Earth observation (EO) classification is essential, as it can enhance semantic segmentation performance and help prevent further exploitation of the results in the case of erroneous prediction. The model we developed, Confidence Assessment for enhanced Semantic segmentation (CAS), evaluates confidence at both the segment and pixel levels, providing both labels and confidence scores as output. Our model, CAS, identifies segments with incorrectly predicted labels using the proposed combined confidence metric, refines the model, and enhances its performance. This work has significant applications, particularly in evaluating EO Foundation Models on semantic segmentation downstream tasks, such as land-cover classification using Sentinel-2 satellite data. The evaluation results show that this strategy is effective and that the proposed model CAS outperforms other baseline models.","PeriodicalId":91017,"journal":{"name":"IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society","volume":"22 ","pages":"1-5"},"PeriodicalIF":0.0,"publicationDate":"2024-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10759697","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142844512","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Deadwood is a critical component of forest ecosystems, storing nutrients for plants and serving as a carbon store and emission source. Climate change influences forest ecosystem dynamics with the potential for deadwood to emit carbon more rapidly due to accelerated decay and increased wildfires and increased inputs via mass forest mortality and disturbance events. To objectively inform our understanding of wildfires and associated carbon emissions, this study estimates the carbon content of dead fine woody debris (FWD) using multimodal data, such as Landsat-8 multispectral imagery, Sentinel-1 (C-band) and PALSAR (L-band) synthetic aperture radar (SAR) imagery, and terrain features to estimate the FWD of less than 0.25 in (1 h), 0.25–1 in (10 h), and 1–3 in (100 h). This data fusion provides spectral information to assess vegetation health that correlates with deadwood, as well as penetrability from SAR, resulting in structural information and biomass sensitivity. An ensemble machine learning (ML) model was trained using measurements from the Forest Inventory and Analysis (FIA) Database. A feature importance analysis was also performed to investigate the importance of input features to the model’s performance. A super learner regression (SLR) model composed of 9 base learners, including an ElasticNet model as meta-learner, was proposed and achieved the $R^{2}$ values of 0.75, 0.72, and 0.62 to estimate 1-, 10-, and 100-h FWD, respectively. The validated model was then used to estimate deadwood carbon in the 2021 Dixie Fire region of California, demonstrating the effectiveness of our approach, emphasizing the value of multimodal data for real-time FWD carbon stock estimation.
{"title":"Remote Sensing and Mapping of Fine Woody Carbon With Satellite Imagery and Super Learner","authors":"Riyaaz Uddien Shaik;Mohamad Alipour;Eric Rowell;Adam Watts;Christopher Woodall;Ertugrul Taciroglu","doi":"10.1109/LGRS.2024.3503585","DOIUrl":"https://doi.org/10.1109/LGRS.2024.3503585","url":null,"abstract":"Deadwood is a critical component of forest ecosystems, storing nutrients for plants and serving as a carbon store and emission source. Climate change influences forest ecosystem dynamics with the potential for deadwood to emit carbon more rapidly due to accelerated decay and increased wildfires and increased inputs via mass forest mortality and disturbance events. To objectively inform our understanding of wildfires and associated carbon emissions, this study estimates the carbon content of dead fine woody debris (FWD) using multimodal data, such as Landsat-8 multispectral imagery, Sentinel-1 (C-band) and PALSAR (L-band) synthetic aperture radar (SAR) imagery, and terrain features to estimate the FWD of less than 0.25 in (1 h), 0.25–1 in (10 h), and 1–3 in (100 h). This data fusion provides spectral information to assess vegetation health that correlates with deadwood, as well as penetrability from SAR, resulting in structural information and biomass sensitivity. An ensemble machine learning (ML) model was trained using measurements from the Forest Inventory and Analysis (FIA) Database. A feature importance analysis was also performed to investigate the importance of input features to the model’s performance. A super learner regression (SLR) model composed of 9 base learners, including an ElasticNet model as meta-learner, was proposed and achieved the \u0000<inline-formula> <tex-math>$R^{2}$ </tex-math></inline-formula>\u0000 values of 0.75, 0.72, and 0.62 to estimate 1-, 10-, and 100-h FWD, respectively. The validated model was then used to estimate deadwood carbon in the 2021 Dixie Fire region of California, demonstrating the effectiveness of our approach, emphasizing the value of multimodal data for real-time FWD carbon stock estimation.","PeriodicalId":91017,"journal":{"name":"IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society","volume":"22 ","pages":"1-5"},"PeriodicalIF":0.0,"publicationDate":"2024-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10759699","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142777592","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-19DOI: 10.1109/LGRS.2024.3501407
Francisco Rodrigues do Amaral;Tin Nguyen Trung;Thierry Pellarin;Nicolas Gratiot
The surface water and ocean topography (SWOT) mission offers groundbreaking opportunities to observe fine-scale spatial changes in low elevation coastal zones (LECZs). This study explores the first SWOT data in one of the mission’s most challenging environments: a data-scarce, tropical region with flat topography, heavy urbanization, and a river with a weak, tidally influenced slope. We focus on SWOT’s water surface elevation (WSE) and water surface slope (WSS) products at the reach level, comparing the measurements to in situ data. Our analysis shows that about half of SWOT’s WSE and WSS measurements fall within the desired error budgets, though WSS lacks linear correlation with in situ data. At this early stage, both SWOT’s WSE and WSS require validation in such complex areas. However, as SWOT’s high-resolution observations improve over time and are integrated with other data, they are expected to provide valuable insights into dynamic river and estuarine processes.
{"title":"Challenging SWOT: Early Assessment of Level 2 High-Rate River Products in an Urbanized, Low Elevation Coastal Zone","authors":"Francisco Rodrigues do Amaral;Tin Nguyen Trung;Thierry Pellarin;Nicolas Gratiot","doi":"10.1109/LGRS.2024.3501407","DOIUrl":"https://doi.org/10.1109/LGRS.2024.3501407","url":null,"abstract":"The surface water and ocean topography (SWOT) mission offers groundbreaking opportunities to observe fine-scale spatial changes in low elevation coastal zones (LECZs). This study explores the first SWOT data in one of the mission’s most challenging environments: a data-scarce, tropical region with flat topography, heavy urbanization, and a river with a weak, tidally influenced slope. We focus on SWOT’s water surface elevation (WSE) and water surface slope (WSS) products at the reach level, comparing the measurements to in situ data. Our analysis shows that about half of SWOT’s WSE and WSS measurements fall within the desired error budgets, though WSS lacks linear correlation with in situ data. At this early stage, both SWOT’s WSE and WSS require validation in such complex areas. However, as SWOT’s high-resolution observations improve over time and are integrated with other data, they are expected to provide valuable insights into dynamic river and estuarine processes.","PeriodicalId":91017,"journal":{"name":"IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society","volume":"22 ","pages":"1-5"},"PeriodicalIF":0.0,"publicationDate":"2024-11-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10758312","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142821204","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-18DOI: 10.1109/LGRS.2024.3500021
Runbo Dong;Huadong Guo;Mengxiong Zhou;Hanlin Ye;Guang Liu
The geometric characteristics of Moon-based Earth observation platforms differ significantly from those of satellite platforms, with geolocation being a key factor that impacts data quality. The geolocation of a Moon-based sensor is influenced by three key factors: lunar ephemeris (lunar position and libration), Earth orientation parameters (EOPs), and the Earth reference model. Measurement errors from these three sources can significantly affect the geolocation accuracy of a Moon-based sensor. This study proposes a new unbiased estimation method to quantify the geolocation uncertainty introduced by these factors, based on the fusion of multiversion datasets. The method avoids making assumptions about the error distribution of ephemeris parameters while providing an effective approximation of the spatiotemporal patterns of geolocation uncertainty. We integrate three types of ephemeris data, three Earth reference models, and multiple EOPs datasets to assess the overall distribution of geolocation uncertainty and separately evaluate the geolocation uncertainty introduced by each individual factor using control variates method. The results indicate that the maximum total geolocation uncertainty caused by the three factors is about 46 m. Ephemeris errors are the dominant contributor, accounting for more than 98% of the total uncertainty. In addition, measurement errors in lunar libration also account for why longitudinal uncertainty is significantly greater than latitudinal uncertainty.
{"title":"Geolocation Uncertainty Analysis of Moon-Based Earth Observations","authors":"Runbo Dong;Huadong Guo;Mengxiong Zhou;Hanlin Ye;Guang Liu","doi":"10.1109/LGRS.2024.3500021","DOIUrl":"https://doi.org/10.1109/LGRS.2024.3500021","url":null,"abstract":"The geometric characteristics of Moon-based Earth observation platforms differ significantly from those of satellite platforms, with geolocation being a key factor that impacts data quality. The geolocation of a Moon-based sensor is influenced by three key factors: lunar ephemeris (lunar position and libration), Earth orientation parameters (EOPs), and the Earth reference model. Measurement errors from these three sources can significantly affect the geolocation accuracy of a Moon-based sensor. This study proposes a new unbiased estimation method to quantify the geolocation uncertainty introduced by these factors, based on the fusion of multiversion datasets. The method avoids making assumptions about the error distribution of ephemeris parameters while providing an effective approximation of the spatiotemporal patterns of geolocation uncertainty. We integrate three types of ephemeris data, three Earth reference models, and multiple EOPs datasets to assess the overall distribution of geolocation uncertainty and separately evaluate the geolocation uncertainty introduced by each individual factor using control variates method. The results indicate that the maximum total geolocation uncertainty caused by the three factors is about 46 m. Ephemeris errors are the dominant contributor, accounting for more than 98% of the total uncertainty. In addition, measurement errors in lunar libration also account for why longitudinal uncertainty is significantly greater than latitudinal uncertainty.","PeriodicalId":91017,"journal":{"name":"IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society","volume":"22 ","pages":"1-5"},"PeriodicalIF":0.0,"publicationDate":"2024-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142810199","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-15DOI: 10.1109/LGRS.2024.3496482
Yifan Ma;Xiaotao Wen;Yang Lei;Wu Wen;Hongping Ren
In oilfield exploration and development, drilling noise creates significant interference, severely reducing the signal-to-noise ratio (SNR) of seismic data. Due to the complex characteristics of noise and signal, suppressing noise while recovering effective signals poses a challenge for denoising models. To address this, we propose a multiscale generator network with adaptive feature extraction for drilling noise suppression. The constructed network primarily consists of a dual-layer encoder-decoder structure. In the encoder, we designed an adaptive feature extraction module (AFEM) and a depthwise separable encoding module. The former utilizes deformable convolutions for adaptive extraction of effective features in seismic data, while the latter employs depthwise separable convolutions and an inverse bottleneck design to reduce computational complexity while maintaining effective feature extraction. The decoding module in the decoder uses two convolutional layers to reconstruct the seismic data with minimal computational cost. To prevent network degradation, residual connections are employed in both the encoding and decoding modules. The dual-layer structure extracts semantic information at different scales, preserving richer effective signal features and maximizing drilling noise suppression. Experimental results on both synthetic and field data demonstrate that the proposed method achieves higher-quality denoising results compared to denoising convolutional neural network (CNNS) (DnCNN), Unet, and MLGNet, while retaining the maximum amount of effective data.
{"title":"Suppressing Drilling Noise From Seismic Data Based on Multiscale Generator Network With Adaptive Feature Extraction","authors":"Yifan Ma;Xiaotao Wen;Yang Lei;Wu Wen;Hongping Ren","doi":"10.1109/LGRS.2024.3496482","DOIUrl":"https://doi.org/10.1109/LGRS.2024.3496482","url":null,"abstract":"In oilfield exploration and development, drilling noise creates significant interference, severely reducing the signal-to-noise ratio (SNR) of seismic data. Due to the complex characteristics of noise and signal, suppressing noise while recovering effective signals poses a challenge for denoising models. To address this, we propose a multiscale generator network with adaptive feature extraction for drilling noise suppression. The constructed network primarily consists of a dual-layer encoder-decoder structure. In the encoder, we designed an adaptive feature extraction module (AFEM) and a depthwise separable encoding module. The former utilizes deformable convolutions for adaptive extraction of effective features in seismic data, while the latter employs depthwise separable convolutions and an inverse bottleneck design to reduce computational complexity while maintaining effective feature extraction. The decoding module in the decoder uses two convolutional layers to reconstruct the seismic data with minimal computational cost. To prevent network degradation, residual connections are employed in both the encoding and decoding modules. The dual-layer structure extracts semantic information at different scales, preserving richer effective signal features and maximizing drilling noise suppression. Experimental results on both synthetic and field data demonstrate that the proposed method achieves higher-quality denoising results compared to denoising convolutional neural network (CNNS) (DnCNN), Unet, and MLGNet, while retaining the maximum amount of effective data.","PeriodicalId":91017,"journal":{"name":"IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society","volume":"22 ","pages":"1-5"},"PeriodicalIF":0.0,"publicationDate":"2024-11-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142798000","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-13DOI: 10.1109/LGRS.2024.3497345
Yaxin Mu;Luzhao Chen;Xin Liu
Magnetic inversion plays an important role in target classification and recognition. However, traditional magnetic susceptibility inversion methods often fall into local minima, and require high storage space and computing power. This letter proposes a novel 2-D magnetic inversion method for compact ferrous targets, the 3-D object is projected onto a 2-D inversion plane and a linear inversion model based on the modulus of magnetic moment on a 2-D horizontal slice is created. The compact anomaly target is composed of multiple nonzero uniformly connected grids in spatial domain, so the target is sparse in the gradient domain, because the nonzero elements on the inversion plane are only located at the boundary of the target. Using the sparseness, total variation compressed sensing (TV-CS) framework is employed to solve the 2-D magnetic inversion problem, achieving rapid classification and identification of hidden targets. In addition, numerical simulations have been performed, and experimental results indicate that the proposed algorithm is able to accurately delineate targets of different shapes and distinguish between solid and hollow targets. For a compact 3-D target, the 2-D magnetic inversion time is 0.096 s with a spatial resolution of 0.5 m, and 2.36 s with a spatial resolution of 0.1 m.
{"title":"A Sparse 2-D Magnetic Inversion Method for Subsurface Target Classification","authors":"Yaxin Mu;Luzhao Chen;Xin Liu","doi":"10.1109/LGRS.2024.3497345","DOIUrl":"https://doi.org/10.1109/LGRS.2024.3497345","url":null,"abstract":"Magnetic inversion plays an important role in target classification and recognition. However, traditional magnetic susceptibility inversion methods often fall into local minima, and require high storage space and computing power. This letter proposes a novel 2-D magnetic inversion method for compact ferrous targets, the 3-D object is projected onto a 2-D inversion plane and a linear inversion model based on the modulus of magnetic moment on a 2-D horizontal slice is created. The compact anomaly target is composed of multiple nonzero uniformly connected grids in spatial domain, so the target is sparse in the gradient domain, because the nonzero elements on the inversion plane are only located at the boundary of the target. Using the sparseness, total variation compressed sensing (TV-CS) framework is employed to solve the 2-D magnetic inversion problem, achieving rapid classification and identification of hidden targets. In addition, numerical simulations have been performed, and experimental results indicate that the proposed algorithm is able to accurately delineate targets of different shapes and distinguish between solid and hollow targets. For a compact 3-D target, the 2-D magnetic inversion time is 0.096 s with a spatial resolution of 0.5 m, and 2.36 s with a spatial resolution of 0.1 m.","PeriodicalId":91017,"journal":{"name":"IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society","volume":"22 ","pages":"1-5"},"PeriodicalIF":0.0,"publicationDate":"2024-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142691737","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-13DOI: 10.1109/LGRS.2024.3497575
Xin Song;Erhao Gao
In the dynamic field of remote sensing images (RSIs), the challenge of object scale variability and sensor resolution disparities is formidable. Addressing these complexities, we have designed a lightweight remote sensing model named YGNet, tailored for multiscale object detection. It demonstrates excellent performance in detecting both multiscale and small objects within RSIs. The E-RMSK module within YGNet employs a gradient-based architecture with multiple parallel reparameterized convolutions in its internal branches, facilitating the extraction of multiscale features while maintaining parameter and computational efficiency. The HLS-PAN structure integrates feature maps extracted through feature selection, enabling the top layers to relay image information downward to lower levels and the lowest layers to transmit data upward for localization, achieving feature fusion. This synergistic effect of the module design enhances the accuracy of object detection in complex remote sensing scenarios and ensures the model’s feasibility on platforms with limited resources. Rigorous testing on the RSOD and NWPU VHR-10 datasets has proven YGNet’s exceptional capabilities, achieving the mean average precision (mAP) scores of 96.2% and 88.9%, respectively. The model meets the demands for real-time, lightweight, multiscale object detection in remote sensing imagery, making it highly suitable for deployment in resource-constrained environments.
{"title":"YGNet: A Lightweight Object Detection Model for Remote Sensing","authors":"Xin Song;Erhao Gao","doi":"10.1109/LGRS.2024.3497575","DOIUrl":"https://doi.org/10.1109/LGRS.2024.3497575","url":null,"abstract":"In the dynamic field of remote sensing images (RSIs), the challenge of object scale variability and sensor resolution disparities is formidable. Addressing these complexities, we have designed a lightweight remote sensing model named YGNet, tailored for multiscale object detection. It demonstrates excellent performance in detecting both multiscale and small objects within RSIs. The E-RMSK module within YGNet employs a gradient-based architecture with multiple parallel reparameterized convolutions in its internal branches, facilitating the extraction of multiscale features while maintaining parameter and computational efficiency. The HLS-PAN structure integrates feature maps extracted through feature selection, enabling the top layers to relay image information downward to lower levels and the lowest layers to transmit data upward for localization, achieving feature fusion. This synergistic effect of the module design enhances the accuracy of object detection in complex remote sensing scenarios and ensures the model’s feasibility on platforms with limited resources. Rigorous testing on the RSOD and NWPU VHR-10 datasets has proven YGNet’s exceptional capabilities, achieving the mean average precision (mAP) scores of 96.2% and 88.9%, respectively. The model meets the demands for real-time, lightweight, multiscale object detection in remote sensing imagery, making it highly suitable for deployment in resource-constrained environments.","PeriodicalId":91017,"journal":{"name":"IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society","volume":"22 ","pages":"1-5"},"PeriodicalIF":0.0,"publicationDate":"2024-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142810449","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","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：