IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing最新文献

{"title":"Cross-Temporal Remote Sensing Image Change Captioning: A Manifold Mapping and Bayesian Diffusion Approach for Land Use Monitoring","authors":"Qingshan Bai;Xiaohua Wang","doi":"10.1109/JSTARS.2025.3575807","DOIUrl":"https://doi.org/10.1109/JSTARS.2025.3575807","url":null,"abstract":"This study proposes a cross-temporal remote sensing image change captioning (RSICC) model named CTM, which is constructed based on manifold mapping and Bayesian diffusion techniques. The primary objective of CTM is to enhance the accuracy and robustness of captioning changes in multitemporal remote sensing images (RSIs). The model first employs manifold mapping to model illumination variations, reducing the impact of seasonal and lighting factors on image consistency. Subsequently, Bayesian diffusion is introduced to improve the modeling capability of cross-temporal image changes, enhancing robustness against noise and pseudo-changes. In addition, a dual-layer multicoding module is adopted to strengthen temporal feature representation, improving the perception of change regions. Finally, a difference enhancement and dual-attention based image-text captioning strategy is proposed to optimize feature selection and enhance the accuracy and detail of textual descriptions. Experimental results demonstrate that CTM exhibits greater robustness in handling long-span RSIs, effectively mitigating pseudo-changes caused by illumination and seasonal variations. On the LEVIR-CC dataset, CTM achieves a CIDEr score of 138.78, outperforming the best existing method by 7.38 points. On the WHU-CDC dataset, CTM achieves the highest performance in BLEU and METEOR metrics, with a CIDEr score of 153.29, showcasing its outstanding performance in RSICC tasks. Furthermore, visual analysis indicates that CTM accurately captures real change regions while significantly suppressing pseudo-changes, maintaining high descriptive accuracy even in complex environments. This study provides an efficient and precise solution for applications such as land use monitoring, environmental monitoring, and disaster response.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":"18 ","pages":"14406-14415"},"PeriodicalIF":4.7,"publicationDate":"2025-06-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11021286","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144308574","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}

{"title":"IMERG-Like Precipitation Retrieval From Geo-Kompsat-2A Observations Using Conditional Generative Adversarial Networks","authors":"Kyung-Hoon Han;Jaehoon Jeong;Sungwook Hong","doi":"10.1109/JSTARS.2025.3575763","DOIUrl":"https://doi.org/10.1109/JSTARS.2025.3575763","url":null,"abstract":"This study proposes an infrared-to-rain (IR2Rain) model to enhance the accuracy of the geostationary (GEO) weather satellite Geo-Kompsat-2A (GK-2A) rain rate (RR) product. The IR2Rain model is built upon a conditional generative adversarial network, taking GK-2A brightness temperatures as inputs and Integrated MultisatellitE Retrievals for Global Precipitation Measurement (IMERG) Final RRs as target values. To address the distinct physical characteristics and ranges of the input and target datasets, IR2Rain employs preprocessing for normalization and postprocessing for denormalization. The IR2Rain model is developed and validated using the paired input and output datasets collected between September 2019 and December 2022, encompassing a broad region across Asia and Oceania. This study compares the performance of IR2Rain-derived RRs against IMERG RR, GK-2A RR, and Precipitation Estimation from Remotely Sensed Information using Artificial Neural Network dynamic infrared (IR) rain rate-now products. The results demonstrated a probability of detection of 0.607, a critical success index of 0.482, a root-mean-square error of 0.759 mm/h, and a correlation coefficient of 0.671. By combining the high temporal resolution of GEO satellite observations with the reliability of IMERG Final data, the IR2Rain model produces a robust near-real-time IMERG-like precipitation product. Despite smoothing effects and the tendency to underestimate intense rainfall, IR2Rain improves the performance relative to RR products based on the same GK-2A IR data, mitigates the latency encountered in IMERG data generation, and provides timely and accurate precipitation information on intensity and distribution. These products are particularly valuable for operational weather forecasting and public end users in Asia and Oceania, supporting disaster preparedness and hydrological applications.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":"18 ","pages":"14467-14479"},"PeriodicalIF":4.7,"publicationDate":"2025-06-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11021294","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144299250","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}

{"title":"CLPM: A Hybrid Network With Cross-Space Learning and Perception-Driven Mechanism for Long-Tailed Remote Sensing Image Classification","authors":"Lei Zhang;Min Kong;Changfeng Jing;Xing Xing","doi":"10.1109/JSTARS.2025.3575292","DOIUrl":"https://doi.org/10.1109/JSTARS.2025.3575292","url":null,"abstract":"Long-tailed distribution is a common issue in remote sensing image classification (RSIC), and many datasets suffer from severe class imbalance. This imbalance often causes the classifier to focus on the head classes with more samples, neglecting the tail classes. As a result, the precision of the tail classes is reduced, which in turn affects the generalization ability of the classifier. To address this problem, a hybrid network based on cross-space learning and perception-driven mechanism (CLPM) is proposed to improve the classification accuracy of samples from the tail classes. The CLPM network consists of three components. The cross-space representation learning branch is designed to enhance the representation capability of tail-class samples by integrating multiscale and multiregion spatial features. In parallel, the adaptive perception classification branch dynamically adjusts the receptive fields to improve generalization across different resolutions and challenging scenarios. In addition, the CLPM innovatively applies the von Mises-Fisher (vMF) distribution to remote sensing images for high-dimensional interclass feature modeling. Building on this, a vMF-based contrastive loss function is proposed. This approach effectively coordinates the learning processes of head and tail classes while enhancing the precision of feature representation. The effectiveness of CLPM is validated on datasets with varying balance ratios, including SIRI-WHU, CLRS, and NWPU-RESISC45. Results show that CLPM significantly improves tail classes accuracy while maintaining high recognition rates for head and middle classes. Compared with the existing methods, CLPM has significant advantages in the overall recognition accuracy, the long-tailed problem, and diversity adaptation.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":"18 ","pages":"14272-14290"},"PeriodicalIF":4.7,"publicationDate":"2025-06-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11018338","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144281308","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}

{"title":"Investigating the Potential of Deep Learning Approaches in the Reconstruction of VNIR-SWIR Hyperspectral Data From Multispectral Imagery","authors":"Michael Alibani;Martina Pastorino;Gabriele Moser;Nicola Acito","doi":"10.1109/JSTARS.2025.3575518","DOIUrl":"https://doi.org/10.1109/JSTARS.2025.3575518","url":null,"abstract":"Hyperspectral (HS) satellite data are of considerable importance for applications such as environmental monitoring and precision agriculture, given the richness of the spectral information they contain. However, HS data typically exhibit limited spatial resolution and are less readily available than multispectral (MS) data. This study, which aims to simulate data with high spectral and spatial resolution, explores the use of attention-based spectral reconstruction (SR) techniques, specifically MST++, MIRNet, AWAN, and Restormer, to derive HS data in the visible near infrared (VNIR) and short-wave infrared (SWIR) from MS imagery. High-resolution MS and HS image pairs are generated from AVIRIS-NG aerial data and employed for training procedures, thereby enabling the reconstruction of HS data that closely resembles the original measurements. The results indicate that SR techniques can considerably enhance the utility of existing MS datasets for HS-dependent applications. Such techniques can effectively be employed to synthesize high-resolution HS data from MS inputs, thereby facilitating the potential for developing a comprehensive end-to-end sensor simulator. This is particularly advantageous in the context of simulating data from a mission that has not yet become operational, as exemplified by the PRISMA-2G data, which could be simulated, for example, from Sentinel-2 data.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":"18 ","pages":"14215-14227"},"PeriodicalIF":4.7,"publicationDate":"2025-06-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11020775","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144281211","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}

{"title":"Color Night-Light Remote Sensing Image Fusion With Two-Branch Convolutional Neural Network","authors":"Jie Wang;Yanling Lu;Yuefeng Wang;Jianwu Jiang","doi":"10.1109/JSTARS.2025.3563399","DOIUrl":"https://doi.org/10.1109/JSTARS.2025.3563399","url":null,"abstract":"Night-light remote sensing imagery (NLRSI) effectively reflects urban economic and human activities and has important value in the field of remote sensing. However, the applications are limited by their low spatial resolution. In recent years, multisource remote sensing image fusion has become an important method for enhancing the spatial and spectral resolution of single-image data. To address the low-resolution limitation of NLRSI, this study proposes a multisource remote sensing image fusion framework based on the two-branch convolutional neural network (TbCNN), which fuses Landsat-8 and NPP/VIIRS data to generate high-resolution color night-light remote sensing imagery (CNLRSI). First, TbCNN features a deep two-branch structure for multiscale feature extraction, yielding richer spatial texture features. The framework also integrates a multilevel feature fusion module and a residual learning mechanism, further improving the fusion performance of CNLRSI. Quantitative evaluations demonstrate TbCNNs superiority over other methods, achieving optimal values in objective evaluation metrics. Second, in the built-up area extraction experiment, CNLRSI better identifies the true morphology of urban built-up areas compared with NPP/VIIRS data, reducing the overestimation of central urban areas and the underestimation of suburban areas in NPP/VIIRS data. Finally, the enhanced classification capability of CNLRSI is quantitatively validated through confusion matrix analysis, achieving a higher Kappa coefficient (0.814 versus 0.774) than NPP/VIIRS in urban pixel recognition.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":"18 ","pages":"11892-11907"},"PeriodicalIF":4.7,"publicationDate":"2025-04-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10980344","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144073019","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}

{"title":"Multiscale Spatial-Spectral CNN-Transformer Network for Hyperspectral Image Super-Resolution","authors":"Jiayang Zhang;Hongjia Qu;Junhao Jia;Yaowei Li;Bo Jiang;Xiaoxuan Chen;Jinye Peng","doi":"10.1109/JSTARS.2025.3565840","DOIUrl":"https://doi.org/10.1109/JSTARS.2025.3565840","url":null,"abstract":"Remarkable strides have been made in super-resolution methods based on deep learning for hyperspectral images (HSIs), which are capable of enhancing the spatial resolution. However, these methods predominantly focus on capturing local features using convolutional neural networks (CNNs), neglecting the comprehensive utilization of global spatial-spectral information. To address this limitation, we innovatively propose a multiscale spatial-spectral CNN-transformer network for hyperspectral image super resolution, namely, MSHSR. MSHSR not only applies the local spatial-spectral characteristics but also innovatively facilitates the collaborative exploration and application of spatial details and spectral data globally. Specifically, we first design a multiscale spatial-spectral fusion module, which integrates dilated-convolution parallel branches and a hybrid spectral attention mechanism to address the strong local correlations in HSIs, effectively capturing and fusing multiscale local spatial-spectral information. Furthermore, in order to fully exploit the global contextual consistency in HSIs, we introduce a sparse spectral transformer module. This module processes the previously obtained local spatial-spectral features, thoroughly exploring the elaborate global interrelationship and long-range dependencies among different spectral bands through a coarse-to-fine strategy. Extensive experimental results on three hyperspectral datasets demonstrate the superior performance of our method, outperforming comparison methods both in quantitative metrics and visual performance.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":"18 ","pages":"12116-12132"},"PeriodicalIF":4.7,"publicationDate":"2025-04-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10980410","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144131614","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}

{"title":"Magnifier: A Multigrained Neural Network-Based Architecture for Burned Area Delineation","authors":"Daniele Rege Cambrin;Luca Colomba;Paolo Garza","doi":"10.1109/JSTARS.2025.3565819","DOIUrl":"https://doi.org/10.1109/JSTARS.2025.3565819","url":null,"abstract":"In crisis management and remote sensing, image segmentation plays a crucial role, enabling tasks like disaster response and emergency planning by analyzing visual data. Neural networks are able to analyze satellite acquisitions and determine which areas were affected by a catastrophic event. The problem in their development in this context is the data scarcity and the lack of extensive benchmark datasets, limiting the capabilities of training large neural network models. In this article, we propose a novel methodology, namely Magnifier, to improve segmentation performance with limited data availability. The Magnifier methodology is applicable to any existing encoder–decoder architecture, as it extends a model by merging information at different contextual levels through a dual-encoder approach: a local and global encoder. Magnifier analyzes the input data twice using the dual-encoder approach. In particular, the local and global encoders extract information from the same input at different granularities. This allows Magnifier to extract more information than the other approaches given the same set of input images. Magnifier improves the quality of the results of +2.65% on average intersection over union while leading to a restrained increase in terms of the number of trainable parameters compared to the original model. We evaluated our proposed approach with state-of-the-art burned area segmentation models, demonstrating, on average, comparable or better performances in less than half of the giga floating point operations per second (GFLOPs).","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":"18 ","pages":"12263-12277"},"PeriodicalIF":4.7,"publicationDate":"2025-04-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10980409","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144124041","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}

{"title":"Regional Tropospheric Delay Prediction Model Based on LSTM-Enhanced Encoder Network","authors":"Yuanfang Peng;Chenglin Cai;Zexian Li;Kaihui Lv;Xue Zhang;Yihao Cai","doi":"10.1109/JSTARS.2025.3565569","DOIUrl":"https://doi.org/10.1109/JSTARS.2025.3565569","url":null,"abstract":"Precise modeling of zenith tropospheric delay (ZTD) is essential for real-time high-precision positioning in global navigation satellite systems. Due to the stochastic variability of atmospheric water vapor across different regions, tropospheric delay exhibits strong regional characteristics. Empirical tropospheric delay models built on the reanalysis of meteorological data often show significant accuracy discrepancies across regions, failing to meet the needs for precise regional ZTD forecasting. Deep learning methods excel in learning complex patterns and dependencies from time-series data. Our study utilized ZTD data from 178 Nevada Geodetic Laboratory stations in Australia during 2023 as ground truth values and modeled them using a long short-term memory (LSTM)-enhanced encoder network. This model incorporated both spatial and temporal information as well as correlations with GPT3 ZTD. Predictions were compared with those from GPT3 ZTD, ERA5 ZTD, artificial neural network (ANN) ZTD, general regression neural network (GRNN) ZTD, and LSTM ZTD. The results showed that the LSTM-enhanced encoder ZTD achieved a root-mean-square error (RMSE) of 14.43 mm and a mean bias close to zero, with mean absolute error and mean correlation coefficient of 12.42 mm and 0.95, respectively. The proposed model outperforms the GPT3, ERA5, ANN, GRNN, and LSTM models, with respective RMSE improvements of approximately 62.3%, 12.3%, 61%, 59.9%, and 60% . In addition, we compared the spatial and temporal properties of the proposed model with those of the GPT3 and ERA5 models. The discussion section further analyzed the prediction performance of different neural network approaches under different prediction periods.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":"18 ","pages":"13348-13358"},"PeriodicalIF":4.7,"publicationDate":"2025-04-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10980316","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144206179","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}

{"title":"Hyperspectral Image Few-Shot Classification Based on Spatial–Spectral Information Complementation and Multilatent Domain Generalization","authors":"Qianhao Yu;Yong Wang","doi":"10.1109/JSTARS.2025.3565894","DOIUrl":"https://doi.org/10.1109/JSTARS.2025.3565894","url":null,"abstract":"Hyperspectral image (HSI) few-shot classification aims to classify HSI samples of novel categories with limited training HSI samples of base categories. However, current methods suffer from two issues: first, ignoring the complementary relationship between spatial and spectral information; and second, performance degradation on base categories due to excessive focus on novel categories. This article proposes a spatial–spectral information complementation and multilatent domain generalization-based framework (SIM). Specifically, given samples of base (novel) categories, a spatial–spectral feature extraction network is designed to extract their spatial–spectral features, which includes two steps. First, multiple spatial–spectral information complementation modules (SSICs) are stacked to extract the complementary features with different scales. Note that each SSIC extracts features with spatial and spectral information, and adopts a spatial–spectral information transmission unit to cross-transmit spatial and spectral information between these two types of features, thus achieving information complementation. Second, a multiscale feature fusion module is utilized to calculate the classification influence scores of the multiscale complementary features to perform layer-by-layer feature fusion, thus obtaining spatial–spectral features. Afterward, the spatial–spectral features are fed into a classification head to obtain the classification results. During training, a multilatent domain generalization network (MLDGN) is designed, which iteratively assigns pseudodomain labels to all samples, and calculates the sample discrimination loss. SIM combines the sample discrimination loss with the classification losses for training. Thus, SIM can extract spatial–spectral features with domain invariance, alleviating the performance degradation on base categories. Extensive results on four HSI datasets demonstrate that SIM outperforms state-of-the-art methods.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":"18 ","pages":"13212-13224"},"PeriodicalIF":4.7,"publicationDate":"2025-04-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10980625","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144178864","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}

{"title":"DE-Unet: Dual-Encoder U-Net for Ultra-High Resolution Remote Sensing Image Segmentation","authors":"Ye Liu;Shitao Song;Miaohui Wang;Hao Gao;Jun Liu","doi":"10.1109/JSTARS.2025.3565753","DOIUrl":"https://doi.org/10.1109/JSTARS.2025.3565753","url":null,"abstract":"In recent years, there has been a growing demand for remote sensing image semantic segmentation in various applications. The key to semantic segmentation lies in the ability to globally comprehend the input image. While recent transformer-based methods can effectively capture global contextual information, they suffer from high computational complexity, particularly when it comes to ultra-high resolution (UHR) remote sensing images, it is even more challenging for these methods to achieve a satisfactory balance between accuracy and computation speed. To address these issues, we propose in this article a CNN-based dual-encoder U-Net for effective and efficient UHR image segmentation. Our method incorporates dual encoders into the symmetrical framework of U-Net. The dual encoders endow the network with strong global and local perception capabilities simultaneously, while the U-Net's symmetrical structure guarantees the network's robust decoding ability. Additionally, multipath skip connections ensure ample information exchange between the dual encoders, as well as between the encoders and decoders. Furthermore, we proposes a context-aware modulation fusion module that guides the encoder–encoder and encoder–decoder data fusion through global receptive fields. Experiments conducted on public UHR remote sensing datasets such as the Inria Aerial and DeepGlobe have demonstrated the effectiveness of proposed method. Specifically on the Inria Aerial dataset, our method achieves a 77.42% mIoU which outperforms the baseline (Guo et al., 2022) by 3.14% while maintaining comparable inference speed as shown in Fig. <xref>1</xref>.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":"18 ","pages":"12290-12302"},"PeriodicalIF":4.7,"publicationDate":"2025-04-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10980298","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144125603","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}