IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society最新文献_第7页

A Method for Estimating Winter Wheat Height Using UAV Point Cloud Data Enhanced by Density Consistency Filtering 基于密度一致性滤波增强的无人机点云数据估计冬小麦高度方法

IF 4.4

IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society Pub Date : 2026-01-01 Epub Date: 2026-02-11 DOI: 10.1109/LGRS.2026.3663921

Lu Feng;Minfeng Xing;Haitao Lv;Jiali Shang;Jinfei Wang

{"title":"A Method for Estimating Winter Wheat Height Using UAV Point Cloud Data Enhanced by Density Consistency Filtering","authors":"Lu Feng;Minfeng Xing;Haitao Lv;Jiali Shang;Jinfei Wang","doi":"10.1109/LGRS.2026.3663921","DOIUrl":"https://doi.org/10.1109/LGRS.2026.3663921","url":null,"abstract":"Point clouds generated by structure from motion (SfM) are often affected by significant noise caused by plant movement, as well as missing ground points caused by canopy occlusion. This reduces the quality of canopy and terrain extraction and makes it challenging to accurately estimate the height of crops from uncrewed aerial vehicle (UAV) images. To overcome these limitations, this study introduces a density consistency filtering (DCF) algorithm, which adaptively models local density continuity to distinguish between valid points and noisy points. It effectively preserves local canopy structures while removing clustered noise. Furthermore, a color-spatial interpolation scheme based on ExG-RANSAC is developed to reconstruct missing ground points under dense canopies. Evaluated on six datasets from May to June 2019 covering key stages of winter wheat growth, the method achieved a root mean square error (RMSE) of 7.74 cm, mean absolute error (MAE) of 6.23 cm, and R2 of 0.91. After the early stem elongation stage, the method achieved an RMSE of 4.97 cm. The approach significantly improves the estimation accuracy of the late growth stages, demonstrating strong potential for precision agriculture applications.","PeriodicalId":91017,"journal":{"name":"IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society","volume":"23 ","pages":"1-5"},"PeriodicalIF":4.4,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148491859","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}

引用次数: 0

Synchronous Wind and Wave Retrieval From SWOT Based on Deep Learning 基于深度学习的SWOT同步风浪检索

IF 4.4

IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society Pub Date : 2026-01-01 Epub Date: 2026-06-30 DOI: 10.1109/LGRS.2026.3708684

Mengyu Hao;Weizeng Shao;Yongjun Jia;Xingwei Jiang

{"title":"Synchronous Wind and Wave Retrieval From SWOT Based on Deep Learning","authors":"Mengyu Hao;Weizeng Shao;Yongjun Jia;Xingwei Jiang","doi":"10.1109/LGRS.2026.3708684","DOIUrl":"https://doi.org/10.1109/LGRS.2026.3708684","url":null,"abstract":"The novelty of this letter is to propose a deep learning-based algorithm for synchronous wind and wave retrieval from surface water and ocean topography (SWOT). The Level-2 (L-2) measurements derived from Ka-band radar interferometer (KaRIn) aboard SWOT with a spatial resolution of 250 m are collocated with on-scene observations from >800 global National Data Buoy Center (NDBC) buoys in 2024–2025. To model the relationship between SWOT measurements and ocean conditions, a deep learning approach is developed through the collocations, denoted as Transformer. This approach establishes the dependencies of both wind speed and significant wave height (SWH) on the SWOT-derived variables, i.e., the normalized radar cross section (NRCS) and the volume correlation coefficient. In contrast to the correlation (COR) between wind speed and the volume correlation coefficient, which is less than 0.05, all other absolute values of CORs exceed 0.2. Validation against operational products from Haiyang-2 (HY-2) altimetry yields a 0.69-m root-mean-squared error (RMSE) of SWH with a 0.92 COR and a 0.26 scatter index (SI) by Transformer-based wave algorithm, which is showed closer agreement with HY-2 observations than a 0.81-m RMSE of SWH with a 0.89 COR and a 0.30 SI achieved from the 2-km SWOT L2 product. The wind-wave relation is used as a physical constraint in the Transformer-based wind retrieval algorithm, and the comparison with HY-2 scatterometer shows a 1.25 m/s RMSE with a 0.96 COR and a 0.13 SI, which also shows closer agreement with HY-2 observations than the 1.44-m/s RMSE with a 0.94 COR and a 0.15 SI achieved from the 2-km SWOT L2 product. Our work confirms the robustness of the deep learning-based algorithm for generating SWOT wind and wave products, thereby providing a foundation for the future inversion of wind and wave parameters from China’s next-generation ocean dynamics satellite.","PeriodicalId":91017,"journal":{"name":"IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society","volume":"23 ","pages":"1504105-1504105"},"PeriodicalIF":4.4,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148484290","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}

引用次数: 0

Sat-JEPA-Diff: Caption-Guided Zero-RGB Satellite Image Forecasting via Self-Supervised Diffusion Sat-JEPA-Diff：基于自监督扩散的字幕引导零rgb卫星图像预测

IF 4.4

IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society Pub Date : 2026-01-01 Epub Date: 2026-06-30 DOI: 10.1109/LGRS.2026.3708464

Kürşat Kömürcü;Linas Petkevicius

{"title":"Sat-JEPA-Diff: Caption-Guided Zero-RGB Satellite Image Forecasting via Self-Supervised Diffusion","authors":"Kürşat Kömürcü;Linas Petkevicius","doi":"10.1109/LGRS.2026.3708464","DOIUrl":"https://doi.org/10.1109/LGRS.2026.3708464","url":null,"abstract":"Continuous spatiotemporal forecasting in Earth observation (EO) is severely hindered by atmospheric interference. While generative models excel in modeling transitions, they often rely on optical conditioning. In this letter, we advance our previous Sat-JEPA-Diff architecture, which depended on a <inline-formula> <tex-math>$32~times ~32$ </tex-math></inline-formula> coarse RGB input, by introducing a true zero-RGB temporal steering paradigm. By entirely decoupling semantic temporal reasoning from spatial rendering, we eliminate the need for any past pixel conditioning, bypassing physical sensor blindness. We orchestrate the <inline-formula> <tex-math>$t to t{,}+{,}1$ </tex-math></inline-formula> transition using a dual-stream mechanism: a triple-captioning system that textually drives temporal changes, and a robust latent spatial anchor predicted by an adapted Joint-Embedding Predictive Architecture (I-JEPA) to preserve topological constraints. A novel cross-attention fusion adapter dynamically aligns these linguistic trajectories with the spatial anchor to guide a frozen latent diffusion backbone. Evaluations demonstrate that despite the complete removal of optical inputs, the zero-RGB Sat-JEPA-Diff surpasses its coarse-RGB predecessor in perceptual realism (improving LPIPS from 0.4449 to 0.3781) while maintaining competitive geographical fidelity (GSSIM 0.8974). Code and dataset are available on <uri>https://github.com/VU-AIML/SAT-JEPA-DIFF</uri>","PeriodicalId":91017,"journal":{"name":"IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society","volume":"23 ","pages":"6014405-6014405"},"PeriodicalIF":4.4,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148484808","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}

引用次数: 0

A Gaussian Mixture Model Approach for Unsupervised Detection of Oceanographic Features in the Northwest African Upwelling System 西北非洲上升流系统海洋特征无监督检测的高斯混合模型方法

IF 4.4

IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society Pub Date : 2026-01-01 Epub Date: 2026-03-14 DOI: 10.1109/LGRS.2026.3693649

Fatima Zahrae El Azhary;Anass El Aouni;Khalid Minaoui

引用次数: 0

Negative Backscatter Differential Phases on SHV Radars SHV雷达负后向散射差相

IF 4.4

IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society Pub Date : 2026-01-01 Epub Date: 2026-04-27 DOI: 10.1109/LGRS.2026.3687814

Valery Melnikov;Dušan Zrnić;Michael. Istok

引用次数: 0

Triple-Branch Difference Fusion Network for Land Cover Change Detection With Heterogeneous Remote Sensing Images 异质遥感影像土地覆盖变化检测的三分支差分融合网络

IF 4.4

IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society Pub Date : 2026-01-01 Epub Date: 2026-06-16 DOI: 10.1109/LGRS.2026.3704412

Zhiyong Lv;Yifan Zhang;Jiake Li;Wanting Wang;Cheng Shi

{"title":"Triple-Branch Difference Fusion Network for Land Cover Change Detection With Heterogeneous Remote Sensing Images","authors":"Zhiyong Lv;Yifan Zhang;Jiake Li;Wanting Wang;Cheng Shi","doi":"10.1109/LGRS.2026.3704412","DOIUrl":"https://doi.org/10.1109/LGRS.2026.3704412","url":null,"abstract":"Land cover change detection (LCCD) from multimodal remote sensing images (MRSIs) is crucial for environmental monitoring, disaster response, and urban analysis, but it is still challenging due to the heterogeneous imaging mechanisms of different sensors. In particular, direct pixel-wise comparison across modalities is often unreliable. To address this issue, we propose a triple-branch difference fusion network (TBDFN) for multimodal LCCD. TBDFN is composed of two key components. First, a modality conversion module is designed to alleviate the modality discrepancy via cross-modal transformation, which blends content information from one modality with style characteristics from the other to improve cross-modal consistency. Second, built upon the converted representations, a triple-branch fusion architecture is developed to extract temporal difference features from synthetic aperture radar (SAR) and optical imagery, while the same-time cross-modal difference (ST-CMD) module explicitly models intermodality differences within each temporal phase. The three types of difference features are subsequently aggregated to generate the final change map. Experiments on four datasets, with comparisons against six representative methods, demonstrate that TBDFN achieves superior qualitative and quantitative results, validating its effectiveness and robustness for multimodal LCCD. Our proposed method has, respectively, increased the overall accuracy (OA) rate by 1.29%, 0.17%, 1.07%, and 0.13%. The code of the proposed approach can be available at <uri>https://github.com/ImgSciGroup/2026-TBDFN</uri>","PeriodicalId":91017,"journal":{"name":"IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society","volume":"23 ","pages":"6013405-6013405"},"PeriodicalIF":4.4,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148489273","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}

引用次数: 0

GMTNet: Global Feature Integration Mamba–Transformer Hybrid Network for Infrared Small-Target Detection 红外小目标探测的全局特征集成曼巴-变压器混合网络

IF 4.4

IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society Pub Date : 2026-01-01 Epub Date: 2026-03-12 DOI: 10.1109/LGRS.2026.3673619

Zhengqian Feng;Xiaofeng Li;Yuan Gao;Wang Li;Gang Li;Mingle Zhou

{"title":"GMTNet: Global Feature Integration Mamba–Transformer Hybrid Network for Infrared Small-Target Detection","authors":"Zhengqian Feng;Xiaofeng Li;Yuan Gao;Wang Li;Gang Li;Mingle Zhou","doi":"10.1109/LGRS.2026.3673619","DOIUrl":"https://doi.org/10.1109/LGRS.2026.3673619","url":null,"abstract":"Infrared small-target detection (IRSTD) is crucial for surveillance, rescue, and early warning applications, but it still faces many challenges due to weak signals, limited texture, and cluttered backgrounds. Existing methods based on convolutional neural networks (CNNs) lack global context information, while methods based on Transformers and state-space models may be too complex or too smooth, resulting in missed detections. We propose a Mamba–Transformer hybrid network called GMTNet, which can selectively fuse global semantics and local details. GMTNet contains three key components: 1) a directional state-space Transformer (DiST) module for combining direction-aware long-range modeling with fine local cues; 2) a wavelet-based multidimensional lossless downsampling (WMLD) module to preserve high-frequency information during resolution reduction; and 3) a DualAug module for combining geometric and photometric perturbations for robustness. Experiments on IRSTD-1K, NUAA-SIRST, and NUDT-SIRST show that GMTNet consistently outperforms the state-of-the-art methods and can provide robust and efficient IRSTD.","PeriodicalId":91017,"journal":{"name":"IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society","volume":"23 ","pages":"1-5"},"PeriodicalIF":4.4,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148489417","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}

引用次数: 0

SAFNet: Self-Adaptive Focusing Network for Occluded SAR Target Recognition 基于自适应聚焦网络的SAR目标识别

IF 4.4

IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society Pub Date : 2026-01-01 Epub Date: 2026-06-16 DOI: 10.1109/LGRS.2026.3704434

Pengxiang Qin;Zeyu Wang

{"title":"SAFNet: Self-Adaptive Focusing Network for Occluded SAR Target Recognition","authors":"Pengxiang Qin;Zeyu Wang","doi":"10.1109/LGRS.2026.3704434","DOIUrl":"https://doi.org/10.1109/LGRS.2026.3704434","url":null,"abstract":"Convolutional neural network (CNN)-based methods are widely used for synthetic aperture radar (SAR) image automatic target recognition (ATR). However, their recognition performance degrades markedly under partial occlusion conditions, where key scattering centers are obscured and the inherent spatial structure of targets is disrupted. To address this issue, we propose a self-adaptive focusing network (SAFNet) that leverages a learnable spatial attention mechanism to automatically identify and enhance un occuded, informative regions. The model introduces a self-adaptive focusing mechanism to enhance feature learning on unoccluded and informative target regions, and employs a composite loss function for joint optimization of recognition accuracy and spatial regularization. This design effectively suppresses noise interference and feature fragmentation, while maintaining the spatial structural consistency of the target. Experiments on the MSTAR dataset with 10%–50% synthetic occlusion show that SAFNet achieves robust performance under severe visibility degradation. At 50% occlusion, SAFNet improves recognition accuracy by more than 10% compared to conventional CNN backbones and achieves an approximately 3.5% gain over specialized occluded target recognition.","PeriodicalId":91017,"journal":{"name":"IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society","volume":"23 ","pages":"4010305-4010305"},"PeriodicalIF":4.4,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148489855","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}

引用次数: 0

Stationary Precipitation Area Detection Using the Optical Flow Technique With Weather Radar Images 利用气象雷达图像的光流技术检测静止降水区域

IF 4.4

IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society Pub Date : 2026-01-01 Epub Date: 2026-06-17 DOI: 10.1109/LGRS.2026.3704651

Sun-Jin Mo;Ji-Young Gu;Seungwoo Lee;Bo-Young Ye

{"title":"Stationary Precipitation Area Detection Using the Optical Flow Technique With Weather Radar Images","authors":"Sun-Jin Mo;Ji-Young Gu;Seungwoo Lee;Bo-Young Ye","doi":"10.1109/LGRS.2026.3704651","DOIUrl":"https://doi.org/10.1109/LGRS.2026.3704651","url":null,"abstract":"Early detection of stationary heavy precipitation is critical for mitigating flash flood disasters, yet traditional methods often rely on subjective forecaster skills or simple thresholds, leading to potential bias and inaccuracies. To overcome these limitations, this study proposes a novel algorithm, stationary precipitation area detection using the optical flow technique (SPOT). The key innovation of SPOT is its synergistic integration of precipitation intensity with its motion dynamics. Unlike conventional approaches that analyze these factors separately, SPOT introduces a “slowness factor” derived from optical flow motion vectors. This factor dynamically weights the radar reflectivity, a unique methodology that effectively preserves the hazardous signal from strong, slow-moving precipitation echoes while simultaneously suppressing those from faster-moving systems. By quantifying the combined risk of intensity and system stagnation into a single, objective index, SPOT provides a more physically meaningful basis for hazard detection. To validate its performance, SPOT was applied to two significant heavy rainfall events in South Korea, demonstrating superior performance with an <inline-formula> <tex-math>$F$ </tex-math></inline-formula> 1 score of 0.77 and a critical success index of 0.63, alongside a consistently low false alarm rate. These findings highlight SPOT’s potential as a robust and objective tool for the early detection of hazardous rainfall caused by stationary precipitation.","PeriodicalId":91017,"journal":{"name":"IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society","volume":"23 ","pages":"1001505-1001505"},"PeriodicalIF":4.4,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11569929","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148490322","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}

引用次数: 0

An Improved Geometric and Size Relationship-Based Rotated Object Detection Model for Remote Sensing Images 一种改进的基于几何和尺寸关系的遥感图像旋转目标检测模型

IF 4.4

IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society Pub Date : 2026-01-01 Epub Date: 2026-03-15 DOI: 10.1109/LGRS.2026.3693668

Jianjun Ni;Jiabin Huang;Guangyi Tang;Jie Liu;Weidong Cao;Pengfei Shi

{"title":"An Improved Geometric and Size Relationship-Based Rotated Object Detection Model for Remote Sensing Images","authors":"Jianjun Ni;Jiabin Huang;Guangyi Tang;Jie Liu;Weidong Cao;Pengfei Shi","doi":"10.1109/LGRS.2026.3693668","DOIUrl":"https://doi.org/10.1109/LGRS.2026.3693668","url":null,"abstract":"The object detection in remote sensing images has become a focal point in computer vision research. With the advancement of remote sensing technology, there is an increasing demand for detecting objects at arbitrary orientations. However, objects in remote sensing images often exhibit diverse scales and random rotations. Furthermore, rotated objects typically have high aspect ratios, making bounding box predictions sensitive to angular variations and increasing the complexity of detection. To address these challenges, this letter proposes an improved rotated object detection model based on geometric and size relationships. First, an improved ResNet50 network based on multiscale perception enhancement is proposed, where a feature pyramid network (FPN) and a channel attention mechanism are combined to enhance object features. Then, a coarse-to-fine optimization strategy is employed to precisely detect the rotated object. In the coarse stage, the rotated angle is computed by predicting the geometric relationship between the center points and boundary center points, which reduces the issues of angle periodicity and edge exchangeability. In the refined stage, an adaptive rotated intersection-over-union (IoU) loss is proposed, which incorporates the object’s aspect ratio and area, significantly improving the model’s adaptability across various scales, particularly excelling in high-aspect-ratio objects. Experimental results demonstrate that the proposed model effectively identifies rotated objects and outperforms existing models.","PeriodicalId":91017,"journal":{"name":"IEEE geoscience and remote sensing letters : a publication of the IEEE Geoscience and Remote Sensing Society","volume":"23 ","pages":"6010805-6010805"},"PeriodicalIF":4.4,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148490180","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}

引用次数: 0