ISPRS Journal of Photogrammetry and Remote Sensing最新文献

{"title":"Continuous flood monitoring using on-demand SAR data acquired with different geometries: Methodology and test on COSMO-SkyMed images","authors":"Luca Pulvirenti ,&nbsp;Giuseppe Squicciarino ,&nbsp;Luca Cenci ,&nbsp;Luca Ferraris ,&nbsp;Maria Virelli ,&nbsp;Laura Candela ,&nbsp;Silvia Puca","doi":"10.1016/j.isprsjprs.2025.04.036","DOIUrl":"10.1016/j.isprsjprs.2025.04.036","url":null,"abstract":"<div><div>When large floods occur, satellite data are useful for providing emergency managers with frequent and synoptic maps of affected areas, even on a daily basis. Only the use of on-demand SAR data enables the high-resolution monitoring of flood events through acquisitions performed day and night, and regardless of cloud cover, over different areas. However, continuous flood mapping generally requires combining images acquired with different sensor parameters. In turn, this makes data interpretation and processing quite challenging and might require a time-consuming visual analysis activity, which contrasts with the requirement of fast daily delivery of flood maps to end-users.</div><div>This paper presents a new methodology designed to perform continuous flood monitoring in near real-time using on-demand SAR data. It implements a complete workflow, ranging from satellite tasking and pre-flood reference image collection to flood map generation. The core of the methodology is a new automated algorithm based on change detection that can work with data captured with different imaging geometries. The algorithm is designed to discriminate the change due to the change in the scenario from that due to possible differences in the acquisition parameters of the images. It applies different image processing techniques, such as clustering, histogram equalization, fuzzy logic, and region growing, and implements two electromagnetic models. The algorithm is complemented by a post processing step whose objective is to make the daily flood maps consistent with each other. The methodology was tested on a major flood that hit Italy (Emilia-Romagna region) in May 2023, using COSMO-SkyMed data and benchmark flood maps derived from optical data and from the Rapid Mapping component of the Copernicus Emergency Management Service (CEMS). Additionally, it was applied to another flood event that occurred in Italy (Tuscany region) in November 2023, for which benchmark CEMS products were also available, to further assess its reliability. Across these case studies, the algorithm achieved F1-scores ranging from 76% to 90%, demonstrating that, even when using data acquired with geometries that are non-optimal for flood mapping, the methodology produces reliable results. These results are consistent with those reported in the literature for change detection methods applied to acquisitions from the same orbit and for semi-automated supervised workflows such as those used by CEMS.</div><div>The pseudocode of the algorithm is available at: <span><span>https://github.com/LucaP-CIMA/AUTOWADE2.0-pseudocode</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"225 ","pages":"Pages 382-401"},"PeriodicalIF":10.6,"publicationDate":"2025-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143928748","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"VecMapLocNet: Vision-based UAV localization using vector maps in GNSS-denied environments","authors":"Zhen Wang ,&nbsp;Dianxi Shi ,&nbsp;Chunping Qiu ,&nbsp;Songchang Jin ,&nbsp;Tongyue Li ,&nbsp;Ziteng Qiao ,&nbsp;Yang Chen","doi":"10.1016/j.isprsjprs.2025.04.009","DOIUrl":"10.1016/j.isprsjprs.2025.04.009","url":null,"abstract":"<div><div>Vision-based localization techniques are effective UAV localization solutions for GNSS-denied conditions, however they depend on costly, complex, and seasonally variable satellite images or 3D maps, whereas humans can determine location using vector maps. Inspired by human navigation strategies, we propose VecMapLocNet , which uses vector maps to determine UAV 3-DoF poses (latitude, longitude, and yaw) through cross-modal matching. Three key modules are designed to improve the matching between UAV images and vector maps. The UAV feature extraction module is low-latency and adaptable to various flight altitudes, ensuring it is suitable for airborne deployment. The vector map feature extraction module employs a weighted representation of different map elements, ensuring robustness against changes in visual appearance. Inspired by Fourier transforms, the feature matching module for 3-DoF pose estimation is parameter-free, computationally efficient, and invariant to cross-modal differences. To evaluate VecMapLocNet, we introduce a comprehensive dataset that presents challenges, encompassing seven cities worldwide. Through rigorous experimentation, VecMapLocNet has demonstrated competitive performance compared to existing methods in localization accuracy (84.45% Recall@5 m), yaw estimation (88.61% Recall@5°), and computational efficiency (25.23ms latency on onboard device Jetson Orin). Furthermore, we validated VecMapLocNet’s performance in real-world scenarios, with experimental results confirming its generalization ability, achieving a localization error of 16.7 m and an orientation error of 3.1°. The code and datasets are available at <span><span>https://map.geovisuallocalization.com</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"225 ","pages":"Pages 362-381"},"PeriodicalIF":10.6,"publicationDate":"2025-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143928810","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A new approach to color correction and equalization for generating mars global color image mosaics from Tianwen-1 MoRIC images","authors":"Xin Ren ,&nbsp;Xiaoxia Zhang ,&nbsp;Wangli Chen ,&nbsp;Wei Yan ,&nbsp;Xingguo Zeng ,&nbsp;Xu Tan ,&nbsp;Xingye Gao ,&nbsp;Qiang Fu ,&nbsp;Lin Guo ,&nbsp;Qing Zhang ,&nbsp;Zhaopeng Chen ,&nbsp;Guobin Yu ,&nbsp;Rujin Zhao ,&nbsp;Zhiping He ,&nbsp;Jianjun Liu ,&nbsp;Chunlai Li","doi":"10.1016/j.isprsjprs.2025.04.024","DOIUrl":"10.1016/j.isprsjprs.2025.04.024","url":null,"abstract":"<div><div>Mars is rich in color information, and color images can more accurately represent the surface morphology of Mars, aiding in the interpretation of its geomorphological and geological characteristics. Obtaining complete, consistent, well-quantified, and intuitive colors of Mars from remote sensing is challenging due to the instrument limitations, varying the illumination and atmospheric conditions, the dynamics of active geomorphology. Moderate Resolution Imaging Camera (MoRIC) and Mars Mineralogical Spectrometer (MMS) onboard the Tianwen-1 orbiter achieved the global coverage of Mars stereo images and the global spectral measurements in the visible and near-infrared bands. However, due to the effects of illumination and atmospheric conditions, there are obvious north–south stripes in the MoRIC color mosaic map, and the overall color tone is significantly reddish. The color difference analysis shows that the color variation in the high-latitude region of the Northern Hemisphere is mainly caused by different atmospheric conditions at the time of imaging, while in the middle and low latitude regions, it is primarily due to varying lighting conditions. In the high-latitude region of the Southern Hemisphere, the color difference is mainly caused by changes in features associated with the seasonal melting of the polar cap. In this work, we used MMS spectral data to establish Martian surface standard colors and presented a novel approach of quantified color correction and equalization for generating Mars global color image mosaics from MoRIC images. This method adjusts the brightness and tone of the images under different imaging conditions to achieve visual consistency and maintain the tonal consistency of similar features globally. The majority (80.1 %) of the overlapping areas with color differences are concentrated in ΔE00 ≤ 3.0, which significantly improves the color and brightness inconsistency of the origin. The final global color mosaics map with a spatial resolution of 76 m shows a “terracotta tone,” which we expect for Mars images, and is a good approximation of what the human eye might see.</div></div>","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"225 ","pages":"Pages 291-301"},"PeriodicalIF":10.6,"publicationDate":"2025-05-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143923182","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Open-vocabulary generative vision-language models for creating a large-scale remote sensing change detection dataset","authors":"Yujie Zan,&nbsp;Shunping Ji,&nbsp;Songtao Chao,&nbsp;Muying Luo","doi":"10.1016/j.isprsjprs.2025.04.023","DOIUrl":"10.1016/j.isprsjprs.2025.04.023","url":null,"abstract":"<div><div>The lack of large-scale, high-quality training data remains a significant barrier in modern change detection studies. Existing datasets for change detection are typically limited in volume, suffer from category imbalance, and exhibit low label quality, thus impeding the development of data-driven deep learning methods and their practical applications. Artificial intelligence-generated content (AIGC) technology, particularly diffusion-based generative visual-language models (VLMs), has demonstrated remarkable performance in open-vocabulary text-to-image generation and image editing tasks. This approach can be naturally used in generating diverse change samples in remote sensing images. In this paper, we investigate the application of open-vocabulary generative VLMs, particularly the Stable Diffusion (SD) model, for editing specific local areas within remote sensing images using textual prompts, aiming to generate large-scale, high-quality change detection datasets with minimal manual effort. Our methodology involves several steps. Firstly, we utilize the text-image discriminative model CLIP to learn new remote sensing vocabulary such as building and forest from a bird’s-eye view. Simultaneously, we fine-tune a text-to-image model using comprehensive remote sensing scene classification and semantic segmentation data we collected, resulting in our SD-T2I-RS model. Secondly, we develop an inpainting model (SD-Inpainting-RS) based on the SD-T2I-RS model to enable region-based editing capabilities in remote sensing images. Thirdly, we propose a practical strategy for generating change samples using the fine-tuned SD-Inpainting-RS model, aiming to produce more diverse change scenarios. Finally, we introduce the WHU-GCD dataset, a large-scale, high-quality generative change detection dataset supporting both binary change detection (BCD) and semantic change detection (SCD) tasks across 25 semantic change directions. We demonstrate the advantages of the generative change detection dataset in controlling sample scale, category balance, and label accuracy. Additionally, we show that our generative dataset outperforms existing real-world datasets in training superior learning models in the remote sensing domain for the first time. Moreover, the proposed AIGC-based sample generation pipeline not only enhances change detection but also benefits related applications in remote sensing such as land cover classification and instance segmentation. The models and dataset will be publicly available at <span><span>http://gpcv.whu.edu.cn/data</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"225 ","pages":"Pages 275-290"},"PeriodicalIF":10.6,"publicationDate":"2025-05-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143911726","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"BGSNet: A boundary-guided Siamese multitask network for semantic change detection from high-resolution remote sensing images","authors":"Jiang Long ,&nbsp;Sicong Liu ,&nbsp;Mengmeng Li ,&nbsp;Hang Zhao ,&nbsp;Yanmin Jin","doi":"10.1016/j.isprsjprs.2025.04.030","DOIUrl":"10.1016/j.isprsjprs.2025.04.030","url":null,"abstract":"<div><div>The accurate identification of land-surface changed classes when locating changed areas with regular boundaries from satellite images represents a significant challenge, particularly in these areas with considerable spectra and seasonal differences. This paper develops a boundary-guided Siamese multitask network, namely BGSNet, for the purpose of semantic change detection (SCD) from high-resolution remote sensing images. The objective of BGSNet is to utilize robust boundary semantics to enhance the intra-class consistency of change features, alleviating the pseudo-changes caused by temporal variances while retaining well boundary details. The proposed BGSNet consists of three tasks including bi-temporal semantic segmentation, changed areas detection, and boundary detection tasks. In particular, for the semantic segmentation task, a Siamese multilevel pyramid network based on transformer as feature extractors is introduced to fully capture robust semantic features of bi-temporal remote sensing images. For the boundary detection task, a multi-scale feature decoder is designed to enhance boundary semantic representation. For the change detection task, a boundary-contextual guided module is constructed to supply fine-grained semantic constraints, refining the boundaries of detected areas. Finally, we introduce a multitask self-adaptive weighting loss function that considers task uncertainty, which effectively balances the learning effects of different tasks, and improves the model’s adaptability to varied semantic change scenarios. Extensive experiments were conducted on the JiLin-1 and HRSCD datasets, demonstrating that BGSNet outperformed the eight state-of-the-art methods in identifying various semantic changes. Our methods produced the highest attribute accuracy, exceeding reference methods by 5.41%-29.63% and 2.94%-28.09% in SeK measures. Moreover, the detected results by BGSNet exhibited excellent boundary consistency with ground truth, resulting in the lowest geometric errors GTC of 0.187, 0.308, and 0.321 on the JiLin-1, HRSCD, and Fuzhou datasets, respectively. The proposed method also showed a high application promise in large-scale cropland non-agriculturalization scenarios with significant temporal and spectral variations. The code and data will be made available at <span><span>https://github.com/long123524/BGSNet</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"225 ","pages":"Pages 221-237"},"PeriodicalIF":10.6,"publicationDate":"2025-05-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143903335","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"SSRMF: A sparse spectral reconstruction enhanced matched filter for improving point-source methane emission detection in complex terrain","authors":"Ke Li ,&nbsp;Kaixu Bai ,&nbsp;Peng Fu ,&nbsp;Penglong Jiao ,&nbsp;He Chen ,&nbsp;Xinqing Huang ,&nbsp;Chaoshun Liu ,&nbsp;Ni-Bin Chang","doi":"10.1016/j.isprsjprs.2025.04.034","DOIUrl":"10.1016/j.isprsjprs.2025.04.034","url":null,"abstract":"<div><div>Quantification of anthropogenic methane emissions from coal mines and oil &amp; gas facilities is essential to global methane accounting and management. Matched Filter (MF), a classic method for point-source methane emission detection, was proven less effective over heterogeneous land covers given substantial false positives of methane enhancement retrievals due to large uncertainties in background spectrum estimates. To address this challenge, we developed an enhanced MF method, i.e., Sparse Spectral Reconstruction enhanced Matched Filter (SSRMF), to empower accurate methane emission detection in complex terrain from satellite-based hyperspectral images. Specifically, a spectral matrix decomposition and low-rank reconstruction approach was developed to accurately predict background spectrum for each pixel, rather than applying the regional average spectrum that is commonly used in traditional MF method and its alike. Meanwhile, a spatial continuity regularization term was incorporated in the methane enhancement estimation cost function to ensure spatial coherence of methane plumes. Illustration results with synthetic GaoFen-5 hyperspectral images demonstrate the superior advantages of SSRMF over other MF methods, reducing methane enhancement retrieval errors by 80 % and 2.8 times in scenarios with emission flux below 1,000 kg/h. Also, SSRMF enabled to reduce false positives of methane enhancement by 70 % in complex terrain, effectively suppressing artifacts over land covers resembling the methane spectral absorption curve in shortwave infrared bands. Additionally, SSRMF operates 20 % faster than other MF methods by avoiding an iterative optimization of the background spectrum. The error analysis results demonstrate the high robustness of SSRMF against drastic variations in land albedo, aerosol loading, and water vapor, making it more flexible to heterogeneous terrains. By applying SSRMF to 97 clear-sky GaoFen-5A/B hyperspectral images observed during 2019–2022, 126 methane plumes were successfully detected in Shanxi Province, China. Yangquan, Changzhi, and Jincheng were three major cities with notable methane emissions, by a mean flux rate of 6,076 kg/h, 5,470 kg/h, and 4,797 kg/h, respectively. Overall, our proposed SSRMF method provides a more robust solution for methane emission detection in complex terrain from satellite-based hyperspectral images, and can be easily adapted to other trace gases.</div></div>","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"225 ","pages":"Pages 238-256"},"PeriodicalIF":10.6,"publicationDate":"2025-05-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143903336","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"PSeqNet: A crop phenology monitoring model accounting for phenological associations","authors":"Qiyu Tian ,&nbsp;Hao Jiang ,&nbsp;Renhai Zhong ,&nbsp;Xingguo Xiong ,&nbsp;Xuhui Wang ,&nbsp;Jingfeng Huang ,&nbsp;Zhenhong Du ,&nbsp;Tao Lin","doi":"10.1016/j.isprsjprs.2025.04.039","DOIUrl":"10.1016/j.isprsjprs.2025.04.039","url":null,"abstract":"<div><div>Variations of crop phenology are critical indicators of growth conditions and are essential for scheduling irrigation and fertilization activities to mitigate climate risks. Accurately characterizing carry-over climate impacts and phenological associations, especially the delayed influence of earlier stages development on later stages, is key to understanding crop phenological dynamics under changing climate. However, current remote sensing methods face challenges in matching extracted phenological metrics to crop phenological stages and in exploring complex climate interactions. To address these challenges, we propose a novel data-driven phenology monitoring algorithm named Phenology Seq2Seq Network (PSeqNet) to account for underlying phenological associations using fused remote sensing and meteorological data. A two-stream encoder processes and fuses temporal changes of remotely sensed and meteorological information during the growing season, followed by an autoregressive phenological decoder that utilizes the hierarchical structure of phenological development to learn associations among stages. PSeqNet is applied in Northeastern China as a case study to detect and forecast multiple rice phenological stages at the station level. The results indicate that PSeqNet with a two-stream encoder effectively utilizes fused information and extracts distinct associations among stages through its autoregressive decoder. PSeqNet outperformed generic curve fitting and shape model fitting methods in terms of overall accuracy and the degree of correlation, with overall mean absolute error (MAE) ranging from 3.3 to 4.0 days and correlation coefficient (<em>r)</em> ranging from 0.56 to 0.66. Further analysis highlights that the PSeqNet’s ability to capture unique phenological associations such as the weaker correlation between heading and tillering, and the stronger linear correlation between milking and heading. These distinct associations among stages cannot be adequately characterized by curve-fitting and shape model fitting methods. By utilizing the partial seasonal observations, the PSeqNet model also exhibits an outstanding performance in within-season forecasting in a progressive manner (overall MAE decreased from 4.8 to 4.1 days for maturity). Our findings indicate that the PSeqNet presents a promising approach for representing the phenological associations and provides a flexible approach for integrating multi-source information. This robust phenological monitoring approach holds great potential for identifying crop phenological association patterns and their driving climate factors across broader regions.</div></div>","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"225 ","pages":"Pages 257-274"},"PeriodicalIF":10.6,"publicationDate":"2025-05-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143903337","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"PRISMA imaging for land covers and surface materials composition in urban and rural areas adopting multiple endmember spectral mixture analysis (MESMA)","authors":"Giandomenico De Luca ,&nbsp;Jose Luis Pancorbo ,&nbsp;Federico Carotenuto ,&nbsp;Beniamino Gioli ,&nbsp;Giuseppe Modica ,&nbsp;Lorenzo Genesio","doi":"10.1016/j.isprsjprs.2025.04.038","DOIUrl":"10.1016/j.isprsjprs.2025.04.038","url":null,"abstract":"<div><div>Covers and surface materials composition of urban, <em>peri</em>-urban and rural landscapes is significant information for environmental, climate and human-ecosystems interaction monitoring and modeling, as well as for addressing specific urban planning and improving environmental management. In this study the multiple endmember spectral mixture analysis (MESMA) was exploited to overcome the low spatial resolution and spectral mixture of the hyperspectral (HS) satellite PRISMA (<em>PRecursore IperSpettrale della Missione Applicativa</em>). A multi-level detail large-scale mapping of complex urban and rural fractional composition of land covers and surface materials (LCSM) was carried out. High-resolution airborne data enabled the collection of pure endmembers for each impervious and pervious surface materials, also acting as a reference for assessing resulted sub-pixel fractional covers at the pixel scale. Absolute Errors (AE) have shown that MESMA is very promising for quantifying complex landscape composition at the sub-pixel level from PRISMA HS data (overall AE &lt;=0.282; per-class AE &lt; 0.336, with average values ​​even &lt; 0.1 for some classes). Bias Errors (BE) instead attested that under- and overestimation errors for each class were contained in ±0.25 median values for all three levels of detail (i.e., number of classes) tested. These results demonstrate that the proposed framework integrating MESMA and PRISMA HS is a valuable tool to provide detailed land composition in complex landscapes to support urban planning and enhance environmental sustainability.</div></div>","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"225 ","pages":"Pages 196-220"},"PeriodicalIF":10.6,"publicationDate":"2025-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143903334","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"MT_GAN: A SAR-to-optical image translation method for cloud removal","authors":"Peng Wang ,&nbsp;Yongkang Chen ,&nbsp;Bo Huang ,&nbsp;Daiyin Zhu ,&nbsp;Tongwei Lu ,&nbsp;Mauro Dalla Mura ,&nbsp;Jocelyn Chanussot","doi":"10.1016/j.isprsjprs.2025.04.011","DOIUrl":"10.1016/j.isprsjprs.2025.04.011","url":null,"abstract":"<div><div>Synthetic Aperture Radar (SAR) is an active microwave imaging and earth observation device capable of penetrating through clouds, rain, and fog, enabling it to operate effectively regardless of the weather conditions and throughout the day. However, speckle noise in SAR images can make them difficult to interpret, and optical images are often difficult to observe when they are covered by clouds. Therefore, after preprocessing, SAR images can be directly converted to optical images through end-to-end translation learning without optical images as auxiliary information, improving the interpretability of SAR images and realizing cloud removal. Due to the relatively simple structure design of the existing generator based on residual network, it is not perfect to capture and extract the feature information of the image, and the relationship between the features is not well connected, resulting in the existing SAR-optical translation results are not accurate enough. To mitigate this issue, we propose an image translation method utilizing a multilayer translation generative adversarial network (MT_GAN) for cloud removal. First, we design a despeckling module (DSM) to preprocess the speckle noise in SAR. Furthermore, a multilayer translation generator (MTG) is designed for SAR-to-optical (S-O) image translation. It can perform multi-scale image translation on different layers and combine them to enrich the semantic information of features and optimize the translation results. In addition, MTG combined with PatchGAN discriminator is used to compose the optical image generation sub-network (OGS) and SAR image regression sub-network (SRS). Finally, the SRS and OGS are used to establish the connection of cycle consistency loss and optimize the generated optical image. We prepare four datasets for experiments, two of which are used for image translation experiments and the other two for cloud removal experiments. The findings demonstrate that our proposed approach outperforms existing methods across all evaluation metrics and reaches 28.6140 and 0.7069 in PSNR and SSIM indicators, which surpass MS-GAN (28.3348, 0.6403) and DSen2-CR (28.3472, 0.6857), and effectively removes the cloud. The datasets and codes are available at <span><span>https://github.com/NUAA-RS/MT_GAN</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"225 ","pages":"Pages 180-195"},"PeriodicalIF":10.6,"publicationDate":"2025-05-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143899564","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Estimation of surface all-wave net radiation from MODIS data using deep residual neural network based on limited samples","authors":"Shaopeng Li ,&nbsp;Bo Jiang ,&nbsp;Shunlin Liang ,&nbsp;Xiongxin Xiao ,&nbsp;Jianghai Peng ,&nbsp;Hui Liang ,&nbsp;Jiakun Han ,&nbsp;Xiuwan Yin","doi":"10.1016/j.isprsjprs.2025.04.035","DOIUrl":"10.1016/j.isprsjprs.2025.04.035","url":null,"abstract":"<div><div>Deep learning methods have demonstrated significant success in estimating land surface parameters from satellite data. However, these methods often require large sample sizes for optimal performance, which can be difficult to obtain. This study introduces a novel framework that combines transfer learning (TL) and data augmentation (DA) to improve the performance of a deep learning model, the residual neural network (ResNet), in estimating daily all-wave net radiation (<em>R_{n_daily}</em>) from Moderate Resolution Imaging Spectroradiometer (MODIS) top-of-atmosphere (TOA) observations using limited samples. The framework involves two main steps: first, constructing a robust base model using augmented training samples generated through image rotation in the source region; and second, fine-tuning this base model in target regions with limited local samples. The framework was tested in three regions: the Continental United States (US), Mainland China (MC), and the tropical zone (TR), all with limited ground measurement data. The US was selected as the source region due to its relatively better sample conditions. The results showed that: (1) the ResNet model trained in the US using augmented samples achieved a validated R² of 0.95, RMSE of 14.31, and bias of −0.28 Wm⁻², which outperformed the multi-layer perceptron (MLP) and ResNet models trained with original samples by reducing the validated RMSEs of 2.77 Wm⁻² and 0.80 Wm⁻², respectively; (2) the transferred base model also performed the best in MC and TR, with R² values of 0.86 and 0.66, RMSEs of 22.22 and 25.25 Wm⁻², and biases of 0.22 Wm⁻² and −0.21 Wm⁻², respectively, leading to a decrease in validated RMSE by 3.20, 1.87, and 1.14 Wm⁻² for MC and by 2.32, 1.12, and 0.55 Wm⁻² for TR compared to the MLP and ResNet model trained directly and the ResNet model trained using the augmented samples in these regions, respectively; and (3) the more comprehensive the pre-training sample, the better the framework’s performance in the target domain. However, challenges related to cloud cover and input window size need to be carefully addressed when applying the new framework. Overall, the results highlight the effectiveness of the proposed framework and provide a promising approach for applying deep learning methods with limited samples.</div></div>","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"225 ","pages":"Pages 131-143"},"PeriodicalIF":10.6,"publicationDate":"2025-05-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143895449","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}