[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-2058":3,"related-2058":106},{"id":4,"title":5,"url":6,"summary":7,"summary_zh":8,"content":9,"source_name":10,"source_url":6,"published_at":11,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":15,"score_detail":16,"sources":24,"tags":26,"primary_topic":32,"secondary_topics":37,"classification":49,"search_phrases":77,"slug":80,"view_count":81,"doi":82,"paper":83,"created_at":105},2058,"Machine Learning-Based Prediction of Fall Armyworm (Spodoptera frugiperda) Outbreaks in Maize Production Systems of Northern Nigeria Using Climate and Remote Sensing Data","https:\u002F\u002Fdoi.org\u002F10.58578\u002Fkijst.v3i3.12026","Fall Armyworm (Spodoptera frugiperda) poses a substantial threat to maize production across sub-Saharan Africa, particularly in Nigeria, where climatic variability intensifies the risk of pest outbreaks. This study developed and evaluated machine learning models for predicting Fall Armyworm outbreaks in northern Nigeria using integrated climate and remote sensing variables. A retrospective modeling framework was applied to simulated but biologically constrained datasets covering a six-year period (2019–2024) and comprising temperature, rainfall, relative humidity, Normalized Difference Vegetation Index (NDVI), and vegetation condition indices. Three supervised learning algorithms—Random Forest, Support Vector Machine, and Gradient Boosting—were trained and validated using five-fold cross-validation. Gradient Boosting achieved the highest predictive accuracy at 92.8%, followed by Random Forest at 89.2% and Support Vector Machine at 84.6%. Temperature, relative humidity, and NDVI emerged as the most influential predictors of outbreak occurrence, while the integration of satellite-derived vegetation indices improved overall model performance. These findings demonstrate the potential of combining machine learning with remote sensing data to develop scalable and cost-effective early warning systems for agricultural pest management. However, because the models were developed using simulated data, their predictive validity requires confirmation using field-collected observational data. The proposed framework contributes to data-driven pest surveillance by providing a basis for anticipating outbreaks and supporting timely decision-making in maize production systems.","草地贪夜蛾（Spodoptera frugiperda）对撒哈拉以南非洲的玉米生产构成重大威胁，在尼日利亚尤为突出，因为气候变率加剧了害虫暴发的风险。本研究开发并评估了机器学习模型，利用气候与遥感综合变量预测尼日利亚北部草地贪夜蛾的暴发。研究采用回顾性建模框架，基于模拟但受生物学约束的数据集，覆盖六年时间（2019—2024年），包括温度、降雨量、相对湿度、归一化植被指数（NDVI）及植被状况指数。三种监督学习算法——随机森林、支持向量机和梯度提升——通过五折交叉验证进行训练与验证。梯度提升取得了最高的预测准确率，达92.8%，其次为随机森林（89.2%）和支持向量机（84.6%）。温度、相对湿度和NDVI是暴发发生最具影响力的预测因子，而卫星衍生植被指数的整合提升了模型的整体性能。这些发现表明，将机器学习与遥感数据相结合，有望开发可扩展且成本效益高的农业害虫管理预警系统。然而，由于模型基于模拟数据开发，其预测效度仍需利用实地观测数据加以验证。所提出的框架为数据驱动的害虫监测提供了依据，有助于预判暴发并支持玉米生产系统中的及时决策。",null,"Kwaghe International Journal of Sciences and Technology","2026-09-09T00:00:00Z","论文",10,false,75,{"impact":17,"substance":18,"depth":19,"authority":20,"freshness":21,"relevant":22,"comment":23},18,20,16,12,9,1,"将机器学习与气候及遥感数据结合用于尼日利亚北部玉米草地贪夜蛾暴发预测，方法框架有参考价值，但基于模拟数据、结论可靠性待田间验证，属细分领域研究进展。",[25],{"name":10,"url":6},[27,28,29,30,31],"智慧农业","农业人工智能","玉米","遥感","病虫害预警",{"code":33,"slug":34,"name":29,"path":35,"parent_code":36},"A3","maize","\u002Ftopic\u002Ffield-crops\u002Fmaize","A",[38,44],{"code":39,"slug":40,"name":41,"path":42,"parent_code":43},"I2","agricultural-ai","农业AI与大模型","\u002Ftopic\u002Fdigital-agriculture\u002Fagricultural-ai","I",{"code":45,"slug":46,"name":47,"path":48,"parent_code":43},"I3","remote-sensing","遥感与农情监测","\u002Ftopic\u002Fdigital-agriculture\u002Fremote-sensing",{"status":50,"locked":14,"version":51,"method":52,"evidence":53},"classified","2026-10-22.2","rules",{"candidates":54,"agricultural_evidence":57},[55,61,67,72],{"code":33,"score":56,"supported":57,"title":58,"abstract":59,"tags":60},15,true,[34],[29,34],[29],{"code":39,"score":56,"supported":57,"title":62,"abstract":64,"tags":66},[63],"machine learning",[65,63],"机器学习",[28],{"code":45,"score":56,"supported":57,"title":68,"abstract":70,"tags":71},[69],"remote sensing",[30,69],[30],{"code":73,"score":22,"supported":14,"title":74,"abstract":75,"tags":76},"I1",[],[],[27],[78,79],"农业人工智能 病虫害预警 智慧农业 玉米","农业人工智能 病虫害预警","农业人工智能病虫害预警智慧农业玉米-2058",3,"10.58578\u002Fkijst.v3i3.12026",{"doi":82,"openalex_id":84,"authors":85,"venue":10,"cited_by_count":96,"oa_url":9,"card":97,"direction":103,"ingested_from":104},"W7212022953",[86,88,90,92,94],{"name":87,"orcid":9},"Okwor Jude I.",{"name":89,"orcid":9},"Attamah Chinyere G.",{"name":91,"orcid":9},"Uchendu Christian N.",{"name":93,"orcid":9},"Onyima John O.",{"name":95,"orcid":9},"Idoko Bartholomew",0,{"tldr":98,"method":99,"finding":100,"direction":101,"opportunity":102},"用气候与遥感数据构建机器学习模型预测尼日利亚北部玉米草地贪夜蛾暴发。","随机森林、SVM、梯度提升，结合温度、降雨、湿度、NDVI等模拟数据，五折交叉验","梯度提升准确率最高达92.8%，温度、相对湿度和NDVI是最关键预测因子。","农业人工智能与决策模型","模型基于模拟数据，亟需用田间实测数据验证，并开发可扩展的实时早期预警系统。","农业遥感与作物表型","openalex","2026-09-10T23:30:27.820508Z",{"total":107,"page":22,"page_size":107,"items":108},6,[109,183,271,325,404,501],{"id":110,"title":111,"url":112,"summary":113,"summary_zh":114,"content":9,"source_name":115,"source_url":112,"published_at":116,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":117,"score_detail":118,"sources":122,"tags":124,"primary_topic":126,"secondary_topics":127,"classification":129,"search_phrases":148,"slug":151,"view_count":152,"doi":153,"paper":154,"created_at":182},3973,"A physical-context-conditioned multimodal fusion network for estimating maize canopy nitrogen content using multi-source UAV remote sensing data","https:\u002F\u002Fdoi.org\u002F10.1016\u002Fj.compag.2026.112465","A physical-context-conditioned multimodal fusion network for estimating maize canopy nitrogen content using multi-source UAV remote sensing data。Computers and Electronics in Agriculture","一种基于物理情境条件调节的多模态融合网络，用于利用多源无人机遥感数据估算玉米冠层氮含量。《计算机与电子农业》","Computers and Electronics in Agriculture","2026-10-01T00:00:00Z",81,{"impact":17,"substance":119,"depth":17,"authority":120,"freshness":21,"relevant":22,"comment":121},22,14,"多源无人机遥感与物理上下文多模态融合估算玉米冠层氮素，方法新颖、数据可靠，对精准施肥有参考价值。",[123],{"name":115,"url":112},[27,28,29,30,125],"氮素营养",{"code":33,"slug":34,"name":29,"path":35,"parent_code":36},[128],{"code":45,"slug":46,"name":47,"path":48,"parent_code":43},{"status":50,"locked":14,"version":51,"method":52,"evidence":130},{"candidates":131,"agricultural_evidence":57},[132,136,140,144],{"code":33,"score":56,"supported":57,"title":133,"abstract":134,"tags":135},[34],[29,34],[29],{"code":45,"score":56,"supported":57,"title":137,"abstract":138,"tags":139},[69],[30,69],[30],{"code":73,"score":22,"supported":14,"title":141,"abstract":142,"tags":143},[],[],[27],{"code":39,"score":22,"supported":14,"title":145,"abstract":146,"tags":147},[],[],[28],[149,150],"无人机 遥感 玉米 氮含量","多模态融合 冠层氮素","无人机遥感玉米氮含量-3973",2,"10.1016\u002Fj.compag.2026.112465",{"doi":153,"openalex_id":155,"authors":156,"venue":115,"cited_by_count":96,"oa_url":9,"card":177,"direction":103,"ingested_from":104},"W7215084428",[157,159,161,164,167,169,171,174],{"name":158,"orcid":9},"Wang Baoju",{"name":160,"orcid":9},"Jing Zhao",{"name":162,"orcid":163},"Yu Yan","https:\u002F\u002Forcid.org\u002F0009-0004-4202-6935",{"name":165,"orcid":166},"Junke Zhu","https:\u002F\u002Forcid.org\u002F0009-0005-7778-9614",{"name":168,"orcid":9},"Xinlong Li",{"name":170,"orcid":9},"Huizheng Wang",{"name":172,"orcid":173},"Weiguang Yang","https:\u002F\u002Forcid.org\u002F0000-0002-9862-5839",{"name":175,"orcid":176},"Yubin Lan","https:\u002F\u002Forcid.org\u002F0000-0001-6664-5571",{"tldr":178,"method":179,"finding":180,"direction":103,"opportunity":181},"提出物理上下文条件化多模态融合网络，用多源无人机遥感估算玉米冠层氮含量。","多源无人机遥感数据与物理上下文条件化多模态融合网络。","该网络能提升玉米冠层氮含量估算精度，融合多源数据优于单一模态。","可探索物理先验与多模态融合在不同作物、生育期及跨区域迁移中的泛化能力。","2026-10-02T23:30:02.413294Z",{"id":184,"title":185,"url":186,"summary":187,"summary_zh":188,"content":9,"source_name":189,"source_url":186,"published_at":190,"category":12,"cover_url":9,"hotness":13,"is_selected":57,"score":117,"score_detail":191,"sources":193,"tags":195,"primary_topic":197,"secondary_topics":198,"classification":200,"search_phrases":232,"slug":235,"view_count":236,"doi":237,"paper":238,"created_at":270},3842,"Raspberry Pi Lightweight Deep Learning Models Deployed for Corn Growth Stage Classification Using PhenoCam Images","https:\u002F\u002Fdoi.org\u002F10.3390\u002Frs18193337","Identifying crop growth stages is essential for comprehensive assessment that facilitates timely and precise agronomic management. However, conventional field-based manual scouting is labor-intensive, time-consuming, and lacks scalability. Digital field images with advanced models deployed on portable edge-computing devices offer an alternative solution for real-time crop monitoring and decision support. Therefore, in this study, the standard 8 corn growth stages were defined as 8 classes based on distinct visual canopy appearance patterns observed in PhenoCam (near-surface [proximal] remote sensing network camera) imagery collected from 10 PhenoCam sites. These classes do not correspond directly to standard agronomic vegetation and reproductive growth stages. For modeling, four lightweight deep learning (DL) models, ELiteCrop0, ELiteCrop1, ELiteCrop4, and MobNetCropV2, were developed and evaluated across five image vertical clipping levels (0–40%) using a supercomputer (NDSU CCAST). The optimized model was subsequently deployed on a single-board Raspberry Pi 5 computer for edge inference. Model training accounted for the majority of the total CPU time, exceeding 97%, while the testing times ranged from 0.01–0.12 min, enabling real-time applications. Among the models, ELiteCrop0 achieved the most balanced performance with a confusion-matrix diagonal ratio (CMDR) of 0.93, followed by ELiteCrop1 (CMDR = 0.92). Overall, model performance decreased with increasing vertical clipping; therefore, a moderate image clipping (0–10%) was recommended for improved computational efficiency. Analysis with a supercomputer produced an intrasite (same training sites) accuracy of 0.90–0.93 (Raspberry Pi: 0.78–0.81) and an intersite (unseen test sites) accuracy of 0.48–0.50 (Raspberry Pi: 0.41–0.43), indicating challenges with model generalization. Raspberry Pi processed ≈ 1000 images\u002Fmin at 68 °C, with a PhenoCam:Raspberry Pi processing ratio of 1:2500 at 50% capacity over an 8 h day. Overall, this study presents a scalable and cost-effective solution for real-time corn growth stage monitoring in precision agriculture, which could be extended to other crops.","识别作物生长阶段对于全面评估、促进及时精准的农艺管理至关重要。然而，传统的田间人工巡查劳动强度大、耗时且缺乏可扩展性。部署在便携式边缘计算设备上的数字田间图像与先进模型为实时作物监测和决策支持提供了替代方案。因此，本研究基于从10个PhenoCam（近地面[近端]遥感网络相机）站点采集的影像中观察到的不同冠层视觉外观模式，将标准的8个玉米生长阶段定义为8个类别。这些类别并不直接对应于标准的农艺营养生长和生殖生长阶段。在建模方面，开发了四种轻量级深度学习（DL）模型——ELiteCrop0、ELiteCrop1、ELiteCrop4和MobNetCropV2，并利用超级计算机（NDSU CCAST）在五种图像垂直裁剪水平（0–40%）下进行了评估。随后，优化后的模型被部署到单板Raspberry Pi 5计算机上进行边缘推理。模型训练占总CPU时间的大部分，超过97%，而测试时间为0.01–0.12分钟，支持实时应用。在各模型中，ELiteCrop0实现了最均衡的性能，混淆矩阵对角线比率（CMDR）为0.93，其次是ELiteCrop1（CMDR = 0.92）。总体而言，模型性能随垂直裁剪比例的增加而下降；因此，建议采用适度的图像裁剪（0–10%）以提高计算效率。使用超级计算机的分析得出站内（相同训练站点）准确率为0.90–0.93（Raspberry Pi：0.78–0.81），站间（未见测试站点）准确率为0.48–0.50（Raspberry Pi：0.41–0.43），表明模型泛化面临挑战。Raspberry Pi在68 °C下处理约1000张图像\u002F分钟，在8小时工作日50%容量下，PhenoCam与Raspberry Pi的处理比为1:2500。总体而言，本研究为精准农业中的实时玉米生长阶段监测提供了一种可扩展且经济高效的解决方案，并可扩展至其他作物。","Remote Sensing","2026-09-29T00:00:00Z",{"impact":17,"substance":119,"depth":17,"authority":120,"freshness":21,"relevant":22,"comment":192},"该研究将轻量深度学习模型部署于树莓派实现玉米生育期实时识别，方法新颖、数据扎实，对精准农业边缘监测有参考价值。",[194],{"name":189,"url":186},[27,28,196,29,30],"边缘计算",{"code":33,"slug":34,"name":29,"path":35,"parent_code":36},[199],{"code":39,"slug":40,"name":41,"path":42,"parent_code":43},{"status":50,"locked":14,"version":51,"method":52,"evidence":201},{"candidates":202,"agricultural_evidence":57},[203,210,215,222,226],{"code":39,"score":17,"supported":57,"title":204,"abstract":206,"tags":209},[205],"deep learning",[207,205,208],"深度学习","decision support",[28],{"code":33,"score":56,"supported":57,"title":211,"abstract":213,"tags":214},[212],"corn",[29,212],[29],{"code":73,"score":216,"supported":57,"title":217,"abstract":218,"tags":221},7,[],[219,220],"精准农业","precision agriculture",[27],{"code":45,"score":216,"supported":57,"title":223,"abstract":224,"tags":225},[],[30,69],[30],{"code":227,"score":228,"supported":57,"title":229,"abstract":230,"tags":231},"I5",4,[],[196],[196],[233,234],"Raspberry Pi 玉米 生长阶段","PhenoCam 玉米 深度学习","RaspberryPi玉米生长阶段-3842",5,"10.3390\u002Frs18193337",{"doi":237,"openalex_id":239,"authors":240,"venue":189,"cited_by_count":96,"oa_url":186,"card":265,"direction":103,"ingested_from":104},"W7214787420",[241,244,247,250,253,256,259,262],{"name":242,"orcid":243},"Anne Joice","https:\u002F\u002Forcid.org\u002F0000-0001-7499-9222",{"name":245,"orcid":246},"Humeera Tazeen","https:\u002F\u002Forcid.org\u002F0000-0003-2411-0725",{"name":248,"orcid":249},"Talha Tufaique","https:\u002F\u002Forcid.org\u002F0000-0002-4490-5341",{"name":251,"orcid":252},"Mathala Juliet Gupta","https:\u002F\u002Forcid.org\u002F0000-0002-3886-2429",{"name":254,"orcid":255},"Cannayen Igathinathane","https:\u002F\u002Forcid.org\u002F0000-0001-8884-7959",{"name":257,"orcid":258},"Nitin Rai","https:\u002F\u002Forcid.org\u002F0000-0002-9350-9601",{"name":260,"orcid":261},"Craig W. Whippo","https:\u002F\u002Forcid.org\u002F0000-0002-7626-6919",{"name":263,"orcid":264},"David Walter Archer","https:\u002F\u002Forcid.org\u002F0000-0002-4816-7040",{"tldr":266,"method":267,"finding":268,"direction":103,"opportunity":269},"用PhenoCam图像训练轻量深度学习模型，在树莓派上实时分类玉米8个生长阶段。","基于10个PhenoCam站点图像，开发4种轻量DL模型，在树莓派5上边缘推理。","ELiteCrop0表现最佳（CMDR=0.93），站点内精度0.90-0.93，但站点间泛化差（0","站点间泛化差是主要瓶颈，可研究跨站点迁移学习或域适应以提升模型鲁棒性。","2026-09-30T23:30:29.168194Z",{"id":272,"title":273,"url":274,"summary":275,"summary_zh":276,"content":9,"source_name":277,"source_url":274,"published_at":278,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":279,"score_detail":280,"sources":284,"tags":286,"primary_topic":289,"secondary_topics":290,"classification":292,"search_phrases":309,"slug":312,"view_count":96,"doi":313,"paper":314,"created_at":324},5570,"Evaluation of Counterfactual Explanation Robustness for Computer Vision Agriculture Yield-Forecast Models Under Satellite Cloud-Mask Gaps: An Interpretability Reliability Bound for Cooperative Storage Decisions","https:\u002F\u002Fdoi.org\u002F10.5281\u002Fzenodo.23545079","Counterfactual explanations are increasingly proposed as the interface through which cooperative storage managers interrogate computer-vision yield-forecast models, yet their reliability under the satellite cloud-mask gaps that routinely afflict agricultural remote sensing has not been bounded. The bound is derived from a Lipschitz continuity assumption on the forecast model and a masking-operator perturbation analysis, and it is expressed as a function of the cloud-mask gap fraction, the model's local Lipschitz constant, and the counterfactual search tolerance. A replication-based evaluation of the before-after-control-impact estimator was conducted to characterise the operating characteristics of the design under a declared variance structure (between-cluster SD = 6.0, residual SD = 4.5, intraclass correlation = 0.64), with power estimated across 2000 design cells. The reliability bound degrades non-linearly as the gap fraction rises, and that the intraclass correlation of 0.64 implies that cooperative-level aggregation of counterfactual recourse is necessary to preserve explanation stability. The contribution is a decision-relevant reliability bound that storage cooperatives can use to set a maximum tolerable cloud-mask gap before counterfactual recourse is suspended.","反事实解释日益被提出作为合作式仓储管理者质询计算机视觉产量预测模型的接口，然而在农业遥感中经常出现的卫星云掩膜缺口条件下，其可靠性尚未得到界定。该界基于预测模型上的Lipschitz连续性假设和掩膜算子扰动分析推导得出，并表示为云掩膜缺口比例、模型局部Lipschitz常数和反事实搜索容差的函数。对前后对照影响估计量进行了基于重复的评估，以表征该设计在给定方差结构（组间标准差=6.0，残差标准差=4.5，组内相关系数=0.64）下的运行特征，并在2000个设计单元中估计了检验效能。可靠性界随缺口比例上升呈非线性退化，且组内相关系数0.64意味着需要对反事实救济进行合作社层面的聚合，以保持解释的稳定性。本文的贡献在于提供了一个与决策相关的可靠性界，仓储合作社可据此设定在暂停反事实救济之前可容忍的最大云掩膜缺口。","Zenodo (CERN European Organization for Nuclear Research)","2026-10-21T00:00:00Z",67,{"impact":281,"substance":18,"depth":282,"authority":20,"freshness":13,"relevant":22,"comment":283},8,17,"该论文提出反事实解释在卫星云掩膜缺口下的可靠性边界，方法新颖且结论可用于合作社决策，但属细分方法学研究，产业影响有限。",[285],{"name":277,"url":274},[27,28,287,30,288],"产量预测","可解释性",{"code":39,"slug":40,"name":41,"path":42,"parent_code":43},[291],{"code":45,"slug":46,"name":47,"path":48,"parent_code":43},{"status":50,"locked":14,"version":51,"method":52,"evidence":293},{"candidates":294,"agricultural_evidence":57},[295,301,305],{"code":39,"score":20,"supported":57,"title":296,"abstract":298,"tags":300},[297],"computer vision",[299],"计算机视觉",[28],{"code":45,"score":281,"supported":57,"title":302,"abstract":303,"tags":304},[],[30,69],[287,30],{"code":73,"score":22,"supported":14,"title":306,"abstract":307,"tags":308},[],[],[27],[310,311],"农业人工智能 产量预测 可解释性 智慧农业","农业人工智能 产量预测","农业人工智能产量预测可解释性智慧农业-5570","10.5281\u002Fzenodo.23545079",{"doi":313,"openalex_id":315,"authors":316,"venue":277,"cited_by_count":96,"oa_url":274,"card":319,"direction":103,"ingested_from":104},"W7223114002",[317],{"name":318,"orcid":9},"Kadiatou Barry",{"tldr":320,"method":321,"finding":322,"direction":103,"opportunity":323},"推导卫星云掩膜缺口下反事实解释可靠性的理论界，并评估其对合作社储粮决策的影响。","基于Lipschitz连续性假设与掩膜扰动分析推导可靠性界，并用2000个设计单","缺口比例上升使可靠性非线性下降；ICC=0.64表明需合作社层级聚合以保持解释稳定。","可研究不同作物区云掩膜统计特性下的自适应解释暂停阈值与聚合策略。","2026-10-22T11:20:39.240791Z",{"id":326,"title":327,"url":328,"summary":329,"summary_zh":330,"content":9,"source_name":331,"source_url":328,"published_at":332,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":333,"score_detail":334,"sources":338,"tags":340,"primary_topic":343,"secondary_topics":344,"classification":352,"search_phrases":378,"slug":381,"view_count":96,"doi":382,"paper":383,"created_at":403},5006,"Mapping Soil Salinity in Arid Agricultural Regions Using Multi-Source Remote Sensing and a Long-Range Dependency-Enhanced Deep Learning Framework","https:\u002F\u002Fdoi.org\u002F10.3390\u002Fland15101966","Soil salinity estimation across arid agricultural regions is challenging due to pronounced environmental heterogeneity and complex dependencies among multi-source variables. This study proposes a Long-Range Dependency-Enhanced Network (LRD-Net) for regional-scale soil salinity estimation in multiple agricultural oasis regions using multi-source remote sensing and environmental covariates. A dataset comprising 180 field samples from three agricultural oasis regions in Xinjiang, China, was constructed with 25 covariates derived from spectral, topographic, climatic, and soil property data. LRD-Net consists of three stages: a multi-scale feature fusion backbone based on depthwise separable convolutions and cross-scale interaction; a Long-Range Dependency Enhancement Module (LDEM) with an asymmetric dual-branch architecture that captures forward and reverse-aware dependencies through reverse alignment and cross-directional fusion; and multi-scale attention refinement with a regression head for continuous salinity estimation. Under the region-stratified random-split evaluation, LRD-Net achieved an R2 of 0.905, an RMSE of 1.207 g·kg−1, and an MAE of 0.868 g·kg−1. Performance decreased under more spatially independent validation, with an R2 of 0.874 under spatial cross-validation and LORO test R2 values ranging from 0.739 to 0.801, indicating that prediction across spatially and geographically independent samples is more challenging than region-stratified random-split evaluation. The proposed framework effectively integrates multi-source information and captures complementary long-range dependencies, providing a potential approach for regional-scale soil salinity estimation within the studied arid agricultural oasis environments and supporting soil salinization monitoring and sustainable land management in the study region.","干旱农业区域土壤盐分估算面临显著的环境异质性和多源变量间复杂依赖关系的挑战。本研究提出了一种长程依赖增强网络（LRD-Net），利用多源遥感与环境协变量对多个农业绿洲区域进行区域尺度土壤盐分估算。研究构建了包含中国新疆三个农业绿洲区域180个野外样本的数据集，并衍生出25个协变量，涵盖光谱、地形、气候和土壤属性数据。LRD-Net由三个阶段组成：基于深度可分离卷积和跨尺度交互的多尺度特征融合骨干网络；具有非对称双分支架构的长程依赖增强模块（LDEM），通过反向对齐和跨方向融合捕获前向与反向感知依赖关系；以及多尺度注意力细化与回归头，用于连续盐分估算。在区域分层随机划分评估下，LRD-Net取得了R²为0.905、RMSE为1.207 g·kg⁻¹、MAE为0.868 g·kg⁻¹的性能。在空间独立性更强的验证下性能有所下降，空间交叉验证R²为0.874，留一区域测试R²范围为0.739至0.801，表明跨空间和地理独立样本的预测比区域分层随机划分评估更具挑战性。所提出的框架有效整合了多源信息并捕获了互补的长程依赖关系，为研究区干旱农业绿洲环境下的区域尺度土壤盐分估算提供了潜在方法，并支持研究区土壤盐渍化监测与可持续土地管理。","Land","2026-10-14T00:00:00Z",72,{"impact":20,"substance":335,"depth":282,"authority":336,"freshness":21,"relevant":22,"comment":337},21,13,"方法新颖、数据扎实的遥感土壤盐渍化估算研究，对干旱区盐碱地监测与治理有参考价值，但属细分领域学术进展，公共影响有限。",[339],{"name":331,"url":328},[27,28,30,341,342],"土壤盐渍化","新疆农业",{"code":45,"slug":46,"name":47,"path":48,"parent_code":43},[345,351],{"code":346,"slug":347,"name":348,"path":349,"parent_code":350},"H2","soil","土壤健康与改良","\u002Ftopic\u002Fenvironment\u002Fsoil","H",{"code":39,"slug":40,"name":41,"path":42,"parent_code":43},{"status":50,"locked":14,"version":51,"method":52,"evidence":353},{"candidates":354,"agricultural_evidence":57},[355,359,364,368,374],{"code":45,"score":56,"supported":57,"title":356,"abstract":357,"tags":358},[69],[30,69],[30],{"code":346,"score":120,"supported":57,"title":360,"abstract":361,"tags":363},[347],[362,347],"土壤",[],{"code":39,"score":21,"supported":57,"title":365,"abstract":366,"tags":367},[205],[],[28],{"code":369,"score":81,"supported":57,"title":370,"abstract":371,"tags":373},"H7",[],[372],"干旱",[],{"code":73,"score":22,"supported":14,"title":375,"abstract":376,"tags":377},[],[],[27],[379,380],"新疆 农业绿洲 土壤盐渍化 遥感","LRD-Net 土壤盐分 估算","新疆农业绿洲土壤盐渍化遥感-5006","10.3390\u002Fland15101966",{"doi":382,"openalex_id":384,"authors":385,"venue":331,"cited_by_count":96,"oa_url":328,"card":398,"direction":103,"ingested_from":104},"W7221816183",[386,388,391,393,395],{"name":387,"orcid":9},"Mingjie Qian",{"name":389,"orcid":390},"Liangjun Wang","https:\u002F\u002Forcid.org\u002F0009-0006-8718-8326",{"name":392,"orcid":9},"Peng Wang",{"name":394,"orcid":9},"Hangyuan Liu",{"name":396,"orcid":397},"Weitao Chen","https:\u002F\u002Forcid.org\u002F0000-0002-6272-1618",{"tldr":399,"method":400,"finding":401,"direction":103,"opportunity":402},"提出LRD-Net深度学习框架，融合多源遥感与环境协变量，估算新疆干旱绿洲区土壤盐分。","构建180个样点、25个协变量的数据集，设计含长程依赖增强模块的多尺度网络。","区域分层随机划分下R²达0.905，但空间独立验证R²降至0.739-0.874，跨区域预测更难。","可探索跨区域迁移学习与不确定性量化，提升空间独立场景下盐分制图泛化能力。","2026-10-15T23:30:29.667342Z",{"id":405,"title":406,"url":407,"summary":408,"summary_zh":409,"content":9,"source_name":410,"source_url":407,"published_at":332,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":117,"score_detail":411,"sources":413,"tags":415,"primary_topic":417,"secondary_topics":418,"classification":426,"search_phrases":473,"slug":476,"view_count":22,"doi":477,"paper":478,"created_at":500},4988,"Hierarchical cross-modal spatiotemporal attention network for crop yield prediction in complex Himalayan agroecosystems","https:\u002F\u002Fdoi.org\u002F10.3389\u002Ffpls.2026.1933033","Introduction Accurate crop yield prediction is essential for precision agriculture, climate-resilient farm management, and timely decision-making. However, yield prediction remains challenging because of complex, nonlinear, and multi-scale interactions among crop physiology, soil conditions, weather variability, and management practices. Conventional statistical, machine learning, and deep learning approaches, including convolutional neural networks (CNNs), recurrent neural networks (RNNs), and long short-term memory (LSTM) networks, often have limitations in capturing long-range temporal dependencies, hierarchical spatial relationships, heterogeneous data modalities, and predictive uncertainty. Methods This study proposes a Hierarchical Cross-Modal Spatio-Temporal Attention Network (HCM-STAN) for crop yield prediction using a unified attention-based framework for integrating heterogeneous agro-environmental data. The architecture incorporates four principal components: (i) hierarchical spatio-temporal attention operating across tree, block, orchard, and regional scales; (ii) a cross-modal attention module for adaptive fusion of IoT-based soil sensor observations, weather data, plant physiological measurements, and drone-based multispectral imagery; (iii) a soil-aware attention head that explicitly models soil memory, hysteresis effects, and root-zone dynamics; and (iv) an uncertainty-aware prediction module combining Monte Carlo dropout and quantile regression to generate calibrated prediction intervals. The framework was evaluated using a multi-source dataset collected from Himalayan apple orchards during 2023–2026, comprising 1,247 monitored trees across three commercial orchards. Results The proposed HCM-STAN achieved an R 2 of 0.89 for four-week-ahead crop yield prediction, representing a 15.6% improvement over the CNN-LSTM baseline. Furthermore, zero-shot transfer to mango orchards achieved an R 2 of 0.67, corresponding to approximately 75% of the performance obtained under full training. Discussion The results demonstrate that integrating hierarchical spatio-temporal representation learning, cross-modal attention, soil-aware modelling, and uncertainty estimation can effectively address the heterogeneous and dynamic nature of crop yield prediction in complex mountain agroecosystems. The observed zero-shot transfer performance further indicates the potential of HCM-STAN to generalize across crop systems without crop-specific retraining. Overall, the proposed framework provides a promising foundation for reliable, uncertainty-aware yield forecasting and can support precision agriculture, climate-smart farm management, and data-driven decision-making in heterogeneous agroecosystems.","引言 准确的作物产量预测对于精准农业、气候适应性农场管理以及及时决策至关重要。然而，由于作物生理、土壤条件、天气变异和管理措施之间复杂、非线性且多尺度的相互作用，产量预测仍然具有挑战性。传统的统计方法、机器学习以及深度学习方法，包括卷积神经网络（CNN）、循环神经网络（RNN）和长短期记忆（LSTM）网络，在捕捉长程时间依赖关系、层次化空间关系、异构数据模态以及预测不确定性方面往往存在局限。方法 本研究提出了一种用于作物产量预测的层次化跨模态时空注意力网络（HCM-STAN），采用统一的基于注意力的框架来整合异构农业环境数据。该架构包含四个主要组成部分：（i）在树、地块、果园和区域尺度上运行的层次化时空注意力；（ii）用于自适应融合基于物联网的土壤传感器观测、天气数据、植物生理测量和无人机多光谱影像的跨模态注意力模块；（iii）显式建模土壤记忆、滞后效应和根区动态的土壤感知注意力头；（iv）结合蒙特卡洛 dropout 和分位数回归以生成校准预测区间的不确定性感知预测模块。该框架使用2023—2026年在喜马拉雅苹果园采集的多源数据集进行了评估，该数据集包含三个商业果园中1，247棵监测树木。结果 所提出的 HCM-STAN 在提前四周作物产量预测中达到了0.89的 R²，相较于 CNN-LSTM 基线提升了15.6%。此外，零样本迁移至芒果园达到了0.67的 R²，相当于完整训练下所获性能的约75%。讨论 结果表明，整合层次化时空表征学习、跨模态注意力、土壤感知建模和不确定性估计，能够有效应对复杂山地农业生态系统中作物产量预测的异构性和动态性。观察到的零样本迁移性能进一步表明，HCM-STAN 具有在无需作物特定再训练的情况下跨作物系统泛化的潜力。总体而言，该 p","Frontiers in Plant Science",{"impact":17,"substance":119,"depth":17,"authority":120,"freshness":21,"relevant":22,"comment":412},"提出分层跨模态时空注意力网络，融合物联网土壤传感、气象、生理与无人机多光谱数据，苹果园四周产量预测R²达0.89并实现芒果园零样本迁移，方法新颖、数据扎实，对智慧农业产量预报有参考价值。",[414],{"name":410,"url":407},[27,28,287,416,30],"多模态融合",{"code":39,"slug":40,"name":41,"path":42,"parent_code":43},[419,425],{"code":420,"slug":421,"name":422,"path":423,"parent_code":424},"B2","fruit","果树","\u002Ftopic\u002Fhorticulture\u002Ffruit","B",{"code":45,"slug":46,"name":47,"path":48,"parent_code":43},{"status":50,"locked":14,"version":51,"method":52,"evidence":427},{"candidates":428,"agricultural_evidence":57},[429,434,442,448,452,456,462,468],{"code":39,"score":430,"supported":57,"title":431,"abstract":432,"tags":433},11,[],[207,65,63,205],[28,416],{"code":420,"score":21,"supported":57,"title":435,"abstract":436,"tags":441},[],[437,438,439,440],"苹果","芒果","orchard","apple",[],{"code":45,"score":281,"supported":57,"title":443,"abstract":444,"tags":447},[],[445,446],"无人机多光谱","multispectral",[287,30],{"code":73,"score":216,"supported":57,"title":449,"abstract":450,"tags":451},[],[219,220],[27],{"code":346,"score":107,"supported":57,"title":453,"abstract":454,"tags":455},[],[362,347],[],{"code":227,"score":107,"supported":57,"title":457,"abstract":458,"tags":461},[],[459,460],"物联网","IoT",[],{"code":463,"score":81,"supported":57,"title":464,"abstract":465,"tags":467},"H6",[],[466],"农业生态",[],{"code":369,"score":81,"supported":57,"title":469,"abstract":470,"tags":472},[],[471],"气候适应",[],[474,475],"喜马拉雅 苹果园 产量预测","HCM-STAN 作物产量预测","喜马拉雅苹果园产量预测-4988","10.3389\u002Ffpls.2026.1933033",{"doi":477,"openalex_id":479,"authors":480,"venue":410,"cited_by_count":96,"oa_url":407,"card":494,"direction":499,"ingested_from":104},"W7221795850",[481,484,486,488,490,492],{"name":482,"orcid":483},"Shabir Ahmad Mir","https:\u002F\u002Forcid.org\u002F0000-0003-2311-218X",{"name":485,"orcid":9},"Umarfarooque Momin",{"name":487,"orcid":9},"Sudeep Kumar E",{"name":489,"orcid":9},"Itfaq Ahmad Mir",{"name":491,"orcid":9},"Ab Rouf Khan",{"name":493,"orcid":9},"Ghulam Mohammad Mir",{"tldr":495,"method":496,"finding":497,"direction":101,"opportunity":498},"提出分层跨模态时空注意力网络HCM-STAN，融合多源农业数据预测喜马拉雅山区苹果产量。","用IoT土壤、气象、生理和无人机多光谱数据，结合分层时空与跨模态注意力及不确定性","四周前产量预测R²达0.89，比CNN-LSTM提升15.6%，零样本迁移芒果R²为0.67。","可探索跨作物零样本迁移的领域自适应机制，并验证模型在更多山地农业生态系统的泛化能力。","智慧农业 \u002F 农业物联网","2026-10-15T23:30:08.750140Z",{"id":502,"title":503,"url":504,"summary":505,"summary_zh":506,"content":9,"source_name":507,"source_url":504,"published_at":508,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":509,"score_detail":510,"sources":512,"tags":514,"primary_topic":516,"secondary_topics":517,"classification":519,"search_phrases":535,"slug":538,"view_count":96,"doi":539,"paper":540,"created_at":554},4759,"Machine Learning-based Prediction of SVI Parameter in Maize: A Comparative Study and Feature Importance Analysis","https:\u002F\u002Fdoi.org\u002F10.24925\u002Fturjaf.v14i10.2870-2879.9043","Background: Accurate estimation of seedling vigor indices (SVI) is critical for monitoring plant health and predicting yields in precision agriculture. This study comparatively evaluates the performance of different machine learning algorithms in estimating the SVI parameter. Methods: Germination data were collected from 39 maize genotypes under three different salinity conditions (0 mM, 100 mM, and 200 mM NaCl), including the parameters Gmax, MGT, GI, SL, RL, SdL, SMR, RMR, and SVI. The dataset was preprocessed through normalization to the [0–1] range. The SVI parameter was estimated using four different machine learning methods: Random Forest, CatBoost, LightGBM, and Support Vector Regression (SVR). To evaluate the performance of the models, both 10-fold cross-validation and 80%–20% hold-out validation techniques were employed. The model performance was examined using RMSE, MSE, MAE, and R² metrics. Results: The findings show that all models have a high explanatory power in the training dataset. Specifically, the CatBoost model showed the highest success in the training phase (R²=0.9962), whereas the SVR model achieved the highest generalization performance on the test dataset (R²=0.9712, RMSE=0.0582). Feature importance analysis results revealed that the 'App' parameter had the highest impact on the target variable in most models, but different structural interactions were prominent in the LightGBM model. The results demonstrate that machine learning techniques provide high accuracy in SVI estimation and that the tendency towards overfitting is limited. Conclusions: In conclusion, this study confirms that machine learning techniques are powerful and reliable tools for estimating spectral indices in precision agriculture applications. The integration of SVR and ensemble-based models into agricultural decision support systems is expected to significantly contribute to optimizing resource use and improving yield predictions.","背景：准确估计幼苗活力指数（SVI）对于精准农业中监测植物健康和预测产量至关重要。本研究比较评估了不同机器学习算法在估计SVI参数方面的性能。方法：从39个玉米基因型在三种不同盐度条件（0 mM、100 mM和200 mM NaCl）下收集发芽数据，包括Gmax、MGT、GI、SL、RL、SdL、SMR、RMR和SVI等参数。数据集通过归一化至[0–1]范围进行预处理。采用四种不同的机器学习方法估计SVI参数：随机森林、CatBoost、LightGBM和支持向量回归（SVR）。为评估模型性能，采用了10折交叉验证和80%–20%留出验证技术。使用RMSE、MSE、MAE和R²指标检验模型性能。结果：研究结果表明，所有模型在训练数据集中均具有较高的解释力。具体而言，CatBoost模型在训练阶段表现出最高的成功率（R²=0.9962），而SVR模型在测试数据集上取得了最高的泛化性能（R²=0.9712，RMSE=0.0582）。特征重要性分析结果显示，在大多数模型中，“App”参数对目标变量的影响最大，但在LightGBM模型中不同的结构交互作用更为突出。结果表明，机器学习技术在SVI估计中提供了高精度，且过拟合倾向有限。结论：总之，本研究证实机器学习技术是精准农业应用中估计光谱指数的强大而可靠的工具。将SVR和基于集成的模型整合到农业决策支持系统中，预计将显著有助于优化资源利用和改善产量预测。","Turkish Journal of Agriculture - Food Science and Technology","2026-10-11T00:00:00Z",63,{"impact":281,"substance":17,"depth":19,"authority":20,"freshness":21,"relevant":22,"comment":511},"方法对比扎实、结论可靠，但属细分领域论文，产业影响有限，适合主题聚合而非头条精选。",[513],{"name":507,"url":504},[27,28,65,29,515],"耐盐育种",{"code":33,"slug":34,"name":29,"path":35,"parent_code":36},[518],{"code":39,"slug":40,"name":41,"path":42,"parent_code":43},{"status":50,"locked":14,"version":51,"method":52,"evidence":520},{"candidates":521,"agricultural_evidence":57},[522,527,531],{"code":39,"score":523,"supported":57,"title":524,"abstract":525,"tags":526},19,[63],[65,63,208],[28,65],{"code":33,"score":56,"supported":57,"title":528,"abstract":529,"tags":530},[34],[29,34],[29],{"code":73,"score":216,"supported":57,"title":532,"abstract":533,"tags":534},[],[219,220],[27],[536,537],"玉米 SVI 机器学习","玉米 耐盐 幼苗活力指数","玉米SVI机器学习-4759","10.24925\u002Fturjaf.v14i10.2870-2879.9043",{"doi":539,"openalex_id":541,"authors":542,"venue":507,"cited_by_count":96,"oa_url":504,"card":549,"direction":101,"ingested_from":104},"W7221462188",[543,546],{"name":544,"orcid":545},"Mustafa Turan Arslan","https:\u002F\u002Forcid.org\u002F0000-0001-5498-6571",{"name":547,"orcid":548},"Gülay Zülkadir","https:\u002F\u002Forcid.org\u002F0000-0003-3488-4011",{"tldr":550,"method":551,"finding":552,"direction":101,"opportunity":553},"用四种机器学习模型基于39个玉米基因型盐胁迫发芽数据预测幼苗活力指数SVI。","39个玉米基因型在三种盐浓度下的发芽数据，归一化后用RF、CatBoost、Li","SVR泛化最优（R²=0.9712），CatBoost训练拟合最高；App参数对SVI影响最大。","可扩展到多作物多胁迫环境，结合光谱或图像表型数据提升SVI预测的田间适用性。","2026-10-12T23:30:44.762497Z"]