[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"topic-LSTM":3},{"name":4,"kind":5,"tokens":6,"total":7,"page":8,"page_size":9,"items":10},"LSTM","tag",[4],2,1,100,[11,91],{"id":12,"title":13,"url":14,"summary":15,"summary_zh":16,"content":17,"source_name":18,"source_url":14,"published_at":19,"category":20,"cover_url":17,"hotness":21,"is_selected":22,"score":23,"score_detail":24,"sources":31,"tags":33,"primary_topic":38,"secondary_topics":44,"classification":45,"search_phrases":58,"slug":61,"view_count":62,"doi":63,"paper":64,"created_at":90},5007,"Kernel-Driven Model Correction of Thermal Directionality: Pixel-Level Analysis Using Airborne Multi-Angular Observations","https:\u002F\u002Fdoi.org\u002F10.3390\u002Frs18203525","Directional anisotropy (DA) in thermal infrared (TIR) remote sensing introduces significant variability in land surface temperature (LST) retrievals, particularly in high-resolution observations. This study presents the empirical pixel-level modelling of DA, by using multi-angular airborne data from the NETSense campaign over an agricultural area near Grosseto, Italy. Six kernel-driven models (KDMs) are evaluated for their ability to fit directional effects, with the Vinnikov–Roujean–Lagouarde (Vinnikov–RL) model achieving the highest accuracy (RMSE = 1.07 K). At 40 m resolution, modelled directional effects reach median deviations of +1.80 K at the hotspot position, −2.03 K opposite to the Sun, and −1.37 K perpendicular to the solar principal plane, with extremes up to ±6 K, highlighting their operational significance for missions such as LSTM, TRISHNA, and SBG. To improve operational feasibility, two simplified models—adapted RL (RLa) and Vinnikov–RLa—are introduced, leveraging a consistent inverse relationship between hotspot magnitude and width to reduce parameter complexity via a power-law formulation. The analysis reveals that directional effects vary with spatial resolution and are influenced by canopy structure and solar geometry, while biophysical parameters alone are insufficient for coefficient estimation. Validation using an independent flight day confirms the robustness of the simplified models under different acquisition geometries, surface types, and geophysical conditions. A tiered model strategy is proposed for future satellite missions, recommending model selection based on day or night observations and data availability. These findings support the integration of directional correction into operational pipelines, enabling consistent and accurate LST products across platforms, resolutions, and observation conditions.","热红外（TIR）遥感中的方向各向异性（DA）会导致地表温度（LST）反演出现显著变化，在高分辨率观测中尤为明显。本研究利用意大利格罗塞托附近农业区NETSense campaign的多角度机载数据，开展了像元尺度的DA经验建模。评估了六种核驱动模型（KDM）拟合方向效应的能力，其中Vinnikov–Roujean–Lagouarde（Vinnikov–RL）模型精度最高（RMSE = 1.07 K）。在40 m分辨率下，模拟的方向效应在热点位置达到中位偏差+1.80 K，背阳方向为−2.03 K，垂直于太阳主平面方向为−1.37 K，极值可达±6 K，凸显了其对LSTM、TRISHNA和SBG等任务的实际意义。为提高业务可行性，引入了两种简化模型——改进RL（RLa）和Vinnikov–RLa，利用热点幅度与宽度之间一致的负相关关系，通过幂律公式降低参数复杂度。分析表明，方向效应随空间分辨率变化，并受冠层结构和太阳几何影响，而仅凭生物物理参数不足以估计系数。利用独立飞行日的验证证实了简化模型在不同观测几何、地表类型和地球物理条件下的稳健性。提出了面向未来卫星任务的分层模型策略，建议根据白天或夜间观测及数据可用性选择模型。这些发现支持将方向校正集成到业务化流程中，从而在不同平台、分辨率和观测条件下实现一致且准确的LST产品。",null,"Remote Sensing","2026-10-14T00:00:00Z","论文",10,false,83,{"impact":25,"substance":26,"depth":27,"authority":28,"freshness":29,"relevant":8,"comment":30},18,23,19,14,9,"基于航空多角度观测的热红外方向性建模研究，方法新颖、数据扎实，对LSTM、TRISHNA等农业遥感卫星任务的地表温度产品校正具有直接参考价值。",[32],{"name":18,"url":14},[4,34,35,36,37],"遥感","地表温度","热红外","TRISHNA",{"code":39,"slug":40,"name":41,"path":42,"parent_code":43},"I3","remote-sensing","遥感与农情监测","\u002Ftopic\u002Fdigital-agriculture\u002Fremote-sensing","I",[],{"status":46,"locked":22,"version":47,"method":48,"evidence":49},"classified","2026-10-22.2","rules",{"candidates":50,"agricultural_evidence":53},[51],{"code":39,"score":52,"supported":53,"title":54,"abstract":55,"tags":57},8,true,[],[34,56],"remote sensing",[35,34],[59,60],"NETSense 航空 多角度 热红外","LSTM TRISHNA SBG 地表温度","NETSense航空多角度热红外-5007",3,"10.3390\u002Frs18203525",{"doi":63,"openalex_id":65,"authors":66,"venue":18,"cited_by_count":82,"oa_url":14,"card":83,"direction":87,"ingested_from":89},"W7221818626",[67,70,73,76,79],{"name":68,"orcid":69},"Louis Snyders","https:\u002F\u002Forcid.org\u002F0000-0003-1093-2517",{"name":71,"orcid":72},"J. A. D. L. Blommaert","https:\u002F\u002Forcid.org\u002F0000-0002-5797-2439",{"name":74,"orcid":75},"J. León-Tavares","https:\u002F\u002Forcid.org\u002F0000-0002-6186-9936",{"name":77,"orcid":78},"Mary Langsdale","https:\u002F\u002Forcid.org\u002F0000-0002-6078-0702",{"name":80,"orcid":81},"Lorenzo Genesio","https:\u002F\u002Forcid.org\u002F0000-0001-9265-886X",0,{"tldr":84,"method":85,"finding":86,"direction":87,"opportunity":88},"基于机载多角度热红外数据，在像素尺度评估核驱动模型并简化，以校正地表温度方向性效应。","使用NETSense机载多角度热红外数据，评估六种核驱动模型并构建简化模型。","Vinnikov–RL模型精度最高，方向性偏差中值达+1.80K至-2.03K，极端达±6K。","农业遥感与作物表型","可探索方向性校正与作物冠层结构参数耦合，提升卫星LST产品在农业监测中的一致性。","openalex","2026-10-15T23:30:29.753962Z",{"id":92,"title":93,"url":94,"summary":95,"summary_zh":17,"content":96,"source_name":97,"source_url":17,"published_at":98,"category":20,"cover_url":17,"hotness":21,"is_selected":22,"score":82,"score_detail":99,"sources":101,"tags":103,"primary_topic":17,"secondary_topics":110,"classification":111,"search_phrases":129,"slug":132,"view_count":29,"doi":17,"paper":17,"created_at":133},73,"Context-Aware Concept Distillation：基于上下文感知的可解释洪涝预测框架（arXiv:2607.23237，2026-07-25提交，预印本）","https:\u002F\u002Farxiv.org\u002Fabs\u002F2607.23237","Levinkopf E、Morin E 与 Goldman CV 在 arXiv 提交预印本\"Context-Aware Concept Distillation for Trustworthy Flood Prediction\"（arXiv:2607.23237 [cs.LG]，2026-07-25提交）。研究面向公共安全高风险场景下深度学习\"黑箱\"阻碍信任与问责的难题，提出 Context-Aware Concept Distillation（CACD）框架，将 LSTM 蒸馏为可解释的水文学感知代理模型：通过无监督流程发现\"水文学语言\"，并用残差超网络（Residual Hypernetwork）根据静态流域特征动态调节这些概念。在全球 5,203 个流域上，模型中位 NSE 0.70，显著优于多层感知机等黑箱基线在未见过未来数据上的表现。","[View PDF](https:\u002F\u002Farxiv.org\u002Fpdf\u002F2607.23237)[HTML (experimental)](https:\u002F\u002Farxiv.org\u002Fhtml\u002F2607.23237v1)\n\n> Abstract:Effective flood risk management relies on accurate forecasting, yet the \"black box\" nature of stateof-the-art Deep Learning models creates a barrier to trust and accountability in high-stakes public safety decisions. While existing Explainable AI (XAI) methods offer local attributions, they fail to provide the verifiable, operationally meaningful causal narratives required by disaster response authorities. To address this societal challenge, we propose Context-Aware Concept Distillation (CACD), a framework developed in collaboration with domain experts to distill opaque LSTMs into interpretable, hydrology-aware surrogate models. We introduce an unsupervised pipeline to discover a \"Hydrological Language\" and a Residual Hypernetwork that dynamically modulates these concepts based on static basin characteristics. Evaluated on 5,203 basins globally, our model achieves high fidelity (Median NSE 0.70), significantly outperforming black-box baselines (e.g., Multi Layer Perceptrons) on unseen future data. By demonstrating that human-interpretable concepts are sufficient to reconstruct flood dynamics, this work balances AI accuracy with the transparency required for responsible environmental decision-making.\n\nComments:to be published in IJCAI 2026 proceedings\nSubjects:Machine Learning (cs.LG); Artificial Intelligence (cs.AI)\nCite as:[arXiv:2607.23237](https:\u002F\u002Farxiv.org\u002Fabs\u002F2607.23237) [cs.LG]\n(or [arXiv:2607.23237v1](https:\u002F\u002Farxiv.org\u002Fabs\u002F2607.23237v1) [cs.LG] for this version)\n[https:\u002F\u002Fdoi.org\u002F10.48550\u002FarXiv.2607.23237](https:\u002F\u002Fdoi.org\u002F10.48550\u002FarXiv.2607.23237)\n\narXiv-issued DOI via DataCite (pending registration)\n\n## Submission history\n\nFrom: Claudia Goldman [[view email](https:\u002F\u002Farxiv.org\u002Fshow-email\u002F51d9175a\u002F2607.23237)] \n\n**[v1]** Sat, 25 Jul 2026 14:45:26 UTC (2,120 KB)","arXiv · 2026-07-25","2026-07-25T00:00:00Z",{"impact":82,"substance":82,"depth":82,"authority":82,"freshness":82,"relevant":82,"comment":100},"论文主题为洪涝预测，不涉及三农、农业信息化、智慧农业、数字乡村或农业科技，不相关。",[102],{"name":97,"url":94},[20,104,105,106,107,108,4,109],"预印本","arXiv","洪涝预测","可解释AI","概念蒸馏","水文学",[],{"status":112,"locked":22,"version":47,"method":48,"evidence":113},"needs_review",{"candidates":114,"agricultural_evidence":22,"reason":128},[115,121],{"code":116,"score":52,"supported":53,"title":117,"abstract":119,"tags":120},"H7",[118],"洪涝",[],[],{"code":122,"score":123,"supported":53,"title":124,"abstract":125,"tags":127},"I2",4,[],[126],"深度学习",[107],"农业相关性需要复核",[130,131],"概念蒸馏 洪涝预测 可解释AI 水文学","概念蒸馏 洪涝预测","概念蒸馏洪涝预测可解释AI水文学-73","2026-07-29T23:54:14.329807Z"]