Pub Date : 2026-02-11DOI: 10.1016/j.isprsjprs.2026.01.041
Qi Li, Lan Zhang, Xi Chen, Chen Zhang, Jingyi Tian, Xianghan Sun, Liqiao Tian
{"title":"Estimation of global riverine total phosphorus concentration based on multi-source data and stacked ensemble learning","authors":"Qi Li, Lan Zhang, Xi Chen, Chen Zhang, Jingyi Tian, Xianghan Sun, Liqiao Tian","doi":"10.1016/j.isprsjprs.2026.01.041","DOIUrl":"https://doi.org/10.1016/j.isprsjprs.2026.01.041","url":null,"abstract":"","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"119 1","pages":""},"PeriodicalIF":12.7,"publicationDate":"2026-02-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146153115","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}
Pub Date : 2026-02-10DOI: 10.1016/j.isprsjprs.2026.02.009
Haoyu Zuo, Minghao Ning, Yiming Shu, Shucheng Huang, Chen Sun
Oriented object detection is critical for enhancing the visual perception of unmanned aerial vehicles (UAVs). However, existing detectors primarily designed for general aerial imagery often struggle to address the unique challenges of UAV imagery, including substantial scale variations, dense clustering, and arbitrary orientations. Furthermore, these models lack probabilistic guarantees required for safety-critical applications. To address these challenges, we propose EAV-DETR, an efficient oriented object detection transformer designed for UAV imagery. Specifically, we first propose a novel scale-adaptive center supervision (SACS) strategy that explicitly enhances the encoder’s feature representations by imposing pixel-level localization constraints with zero inference overhead. Second, we design an anisotropic decoupled rotational attention (ADRA) module, which achieves superior feature alignment for objects of arbitrary morphology by generating a non-rigid adaptive sampling field. Finally, we propose a pose-aware Mondrian conformal prediction (PA-MCP) method, which utilizes the UAV’s flight pose as a physical prior to generate prediction sets with conditional coverage guarantees, thereby providing reliable uncertainty quantification. Extensive experiments on multiple aerial imagery datasets validate the effectiveness of our model. Compared to previous state-of-the-art methods, EAV-DETR improves AP75 on CODrone by 1.76% while achieving a 52% faster inference speed (46.38 vs 30.55 FPS), and improves AP50:95 on UAV-ROD by 3.17%. Our code is available at https://github.com/zzzhak/EAV-DETR.
定向目标检测是提高无人机视觉感知能力的关键。然而，现有的探测器主要是为一般航空图像设计的，通常难以解决无人机图像的独特挑战，包括大量的尺度变化、密集的聚类和任意方向。此外，这些模型缺乏安全关键应用程序所需的概率保证。为了解决这些挑战，我们提出了EAV-DETR，一种针对无人机图像设计的高效定向目标检测转换器。具体来说，我们首先提出了一种新的尺度自适应中心监督（SACS）策略，该策略通过施加零推理开销的像素级定位约束来显式增强编码器的特征表示。其次，我们设计了一个各向异性解耦旋转注意（ADRA）模块，该模块通过生成非刚性自适应采样场来实现对任意形态目标的优越特征对齐。最后，我们提出了一种姿态感知的蒙德里安保形预测（PA-MCP）方法，该方法利用无人机的飞行姿态作为物理先验来生成具有条件覆盖保证的预测集，从而提供可靠的不确定性量化。在多个航空图像数据集上的大量实验验证了我们模型的有效性。与之前最先进的方法相比，EAV-DETR在CODrone上的AP75提高了1.76%，推理速度提高了52% (46.38 vs 30.55 FPS)，在UAV-ROD上的AP50:95提高了3.17%。我们的代码可在https://github.com/zzzhak/EAV-DETR上获得。
{"title":"EAV-DETR: Efficient Arbitrary-View oriented object detection with probabilistic guarantees for UAV imagery","authors":"Haoyu Zuo, Minghao Ning, Yiming Shu, Shucheng Huang, Chen Sun","doi":"10.1016/j.isprsjprs.2026.02.009","DOIUrl":"https://doi.org/10.1016/j.isprsjprs.2026.02.009","url":null,"abstract":"Oriented object detection is critical for enhancing the visual perception of unmanned aerial vehicles (UAVs). However, existing detectors primarily designed for general aerial imagery often struggle to address the unique challenges of UAV imagery, including substantial scale variations, dense clustering, and arbitrary orientations. Furthermore, these models lack probabilistic guarantees required for safety-critical applications. To address these challenges, we propose EAV-DETR, an efficient oriented object detection transformer designed for UAV imagery. Specifically, we first propose a novel scale-adaptive center supervision (SACS) strategy that explicitly enhances the encoder’s feature representations by imposing pixel-level localization constraints with zero inference overhead. Second, we design an anisotropic decoupled rotational attention (ADRA) module, which achieves superior feature alignment for objects of arbitrary morphology by generating a non-rigid adaptive sampling field. Finally, we propose a pose-aware Mondrian conformal prediction (PA-MCP) method, which utilizes the UAV’s flight pose as a physical prior to generate prediction sets with conditional coverage guarantees, thereby providing reliable uncertainty quantification. Extensive experiments on multiple aerial imagery datasets validate the effectiveness of our model. Compared to previous state-of-the-art methods, EAV-DETR improves <mml:math altimg=\"si152.svg\" display=\"inline\"><mml:msub><mml:mrow><mml:mtext>AP</mml:mtext></mml:mrow><mml:mrow><mml:mn>75</mml:mn></mml:mrow></mml:msub></mml:math> on CODrone by 1.76% while achieving a 52% faster inference speed (46.38 vs 30.55 FPS), and improves <mml:math altimg=\"si186.svg\" display=\"inline\"><mml:msub><mml:mrow><mml:mtext>AP</mml:mtext></mml:mrow><mml:mrow><mml:mn>50</mml:mn><mml:mo>:</mml:mo><mml:mn>95</mml:mn></mml:mrow></mml:msub></mml:math> on UAV-ROD by 3.17%. Our code is available at <ce:inter-ref xlink:href=\"https://github.com/zzzhak/EAV-DETR\" xlink:type=\"simple\">https://github.com/zzzhak/EAV-DETR</ce:inter-ref>.","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"9 1","pages":""},"PeriodicalIF":12.7,"publicationDate":"2026-02-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146146708","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}
Pub Date : 2026-02-09DOI: 10.1016/j.isprsjprs.2026.02.006
Yaogang Chen, Jun Hu, Jordi J. Mallorqui, Haiqiang Fu, Wanji Zheng, Aoqing Guo
Interferometric synthetic aperture radar (InSAR) technology can measure ground deformation with high precision over wide areas, which is essential for understanding natural hazards and ensuring infrastructure safety. However, in regions with dense vegetation or frequent surface changes, the radar echoes lose stability over time due to temporal decorrelation. This severely limits the reliability and accuracy of InSAR measurements. Many advanced processing methods have been developed to address this issue, and while they work well in stable conditions, their performance degrades sharply when coherence is lost rapidly. To overcome this limitation, this study proposes a near-real-time sequential multi-temporal polarimetric InSAR (MT-PolInSAR) method tailored for such conditions. For each new acquisition, a stack comprising only the latest images is formed, and statistically homogeneous pixels are reselected dynamically to adapt to evolving scattering mechanisms. A sequential polarimetric-temporal phase optimization is then applied within the stack that confines estimation to short, high-coherence windows and avoids coherence loss between stacks, thereby reducing the effect of fast temporal decorrelation. Deformation time series are subsequently updated through a sequential least squares (LS) inversion using only the newly formed interferograms, which eliminates the need to reprocess the whole dataset and enables timely updates. Experiments with simulated data and full-polarization ALOS-2 and dual-polarization Sentinel-1 images over Fengjie, China, demonstrate that the proposed method significantly increases coherent pixel density and improves deformation accuracy in rapid-decorrelation areas, while enabling genuine near-real-time monitoring with a more efficient processing strategy.
{"title":"A near-real-time multi-temporal polarimetric InSAR method for landslides monitoring in rapid-decorrelation scenarios","authors":"Yaogang Chen, Jun Hu, Jordi J. Mallorqui, Haiqiang Fu, Wanji Zheng, Aoqing Guo","doi":"10.1016/j.isprsjprs.2026.02.006","DOIUrl":"https://doi.org/10.1016/j.isprsjprs.2026.02.006","url":null,"abstract":"Interferometric synthetic aperture radar (InSAR) technology can measure ground deformation with high precision over wide areas, which is essential for understanding natural hazards and ensuring infrastructure safety. However, in regions with dense vegetation or frequent surface changes, the radar echoes lose stability over time due to temporal decorrelation. This severely limits the reliability and accuracy of InSAR measurements. Many advanced processing methods have been developed to address this issue, and while they work well in stable conditions, their performance degrades sharply when coherence is lost rapidly. To overcome this limitation, this study proposes a near-real-time sequential multi-temporal polarimetric InSAR (MT-PolInSAR) method tailored for such conditions. For each new acquisition, a stack comprising only the latest images is formed, and statistically homogeneous pixels are reselected dynamically to adapt to evolving scattering mechanisms. A sequential polarimetric-temporal phase optimization is then applied within the stack that confines estimation to short, high-coherence windows and avoids coherence loss between stacks, thereby reducing the effect of fast temporal decorrelation. Deformation time series are subsequently updated through a sequential least squares (LS) inversion using only the newly formed interferograms, which eliminates the need to reprocess the whole dataset and enables timely updates. Experiments with simulated data and full-polarization ALOS-2 and dual-polarization Sentinel-1 images over Fengjie, China, demonstrate that the proposed method significantly increases coherent pixel density and improves deformation accuracy in rapid-decorrelation areas, while enabling genuine near-real-time monitoring with a more efficient processing strategy.","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"31 1","pages":""},"PeriodicalIF":12.7,"publicationDate":"2026-02-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146146858","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}
Pub Date : 2026-02-06DOI: 10.1016/j.isprsjprs.2026.02.005
Ziyang Xu, Benedikt Schwab, Yihui Yang, Thomas H. Kolbe, Christoph Holst
{"title":"L2M-Reg: Building-level uncertainty-aware registration of outdoor LiDAR point clouds and semantic 3D city models","authors":"Ziyang Xu, Benedikt Schwab, Yihui Yang, Thomas H. Kolbe, Christoph Holst","doi":"10.1016/j.isprsjprs.2026.02.005","DOIUrl":"https://doi.org/10.1016/j.isprsjprs.2026.02.005","url":null,"abstract":"","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"47 1","pages":""},"PeriodicalIF":12.7,"publicationDate":"2026-02-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146135038","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}
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：