[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-2396":3,"related-2396":70},{"id":4,"title":5,"url":6,"summary":7,"summary_zh":8,"content":8,"source_name":9,"source_url":8,"published_at":10,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":14,"score_detail":15,"sources":22,"tags":24,"primary_topic":30,"secondary_topics":36,"classification":37,"search_phrases":54,"slug":57,"view_count":58,"doi":8,"paper":59,"created_at":69},2396,"[预印本]Multi-Task Learning for Sparsely-Labeled Time Series: A Case Study on Cold-Hardiness Modeling","https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.09062","arXiv 2609.09062（2026-09-10）。研究针对葡萄抗寒性的时间序列过程建模问题（数据有限且标签时序稀疏），提出基于RNN的多任务学习方法。实验证明某些MTL架构优于单任务学习和现有科学模型，并在相关萌芽期预测任务上展现良好潜力。",null,"arXiv | 2026-09-10","2026-09-10T00:00:00Z","论文",10,false,66,{"impact":16,"substance":17,"depth":18,"authority":16,"freshness":19,"relevant":20,"comment":21},12,18,16,8,1,"预印本提出RNN多任务学习用于葡萄抗寒性稀疏标签时序建模，方法新颖但尚属早期研究，产业影响有限，可作为农业AI前沿动态收录。",[23],{"name":9,"url":6},[25,26,27,28,29],"农业人工智能","预印本","葡萄","抗寒育种","时间序列",{"code":31,"slug":32,"name":33,"path":34,"parent_code":35},"B2","fruit","果树","\u002Ftopic\u002Fhorticulture\u002Ffruit","B",[],{"status":38,"locked":13,"version":39,"method":40,"evidence":41},"classified","2026-10-22.2","rules",{"candidates":42,"agricultural_evidence":45},[43,49],{"code":31,"score":44,"supported":45,"title":46,"abstract":47,"tags":48},4,true,[],[27],[27],{"code":50,"score":20,"supported":13,"title":51,"abstract":52,"tags":53},"I2",[],[],[25],[55,56],"农业人工智能 抗寒育种 时间序列 预印本","农业人工智能 抗寒育种","农业人工智能抗寒育种时间序列预印本-2396",3,{"doi":8,"openalex_id":8,"authors":60,"venue":8,"cited_by_count":61,"oa_url":8,"card":62,"direction":66,"ingested_from":68},[],0,{"tldr":63,"method":64,"finding":65,"direction":66,"opportunity":67},"提出RNN多任务学习方法，用于标签稀疏的葡萄抗寒时间序列建模。","基于RNN的多任务学习，处理有限且时序稀疏的标签数据。","部分MTL架构优于单任务学习和现有科学模型，并利于萌芽期预测。","农业人工智能与决策模型","可探索多任务学习在更多稀疏标签农业时序场景（如病虫害、物候）中的迁移与泛化。","agent","2026-09-14T00:06:33.014967Z",{"total":71,"page":20,"page_size":71,"items":72},6,[73,158,233,274,348,397],{"id":74,"title":75,"url":76,"summary":77,"summary_zh":78,"content":8,"source_name":79,"source_url":76,"published_at":80,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":14,"score_detail":81,"sources":86,"tags":88,"primary_topic":92,"secondary_topics":93,"classification":100,"search_phrases":134,"slug":137,"view_count":138,"doi":139,"paper":140,"created_at":157},2874,"Laboratory Evaluation of Multispectral Grape Quality Grade Prediction with Compensation for Controlled Illumination Variation","https:\u002F\u002Fdoi.org\u002F10.1088\u002F2631-8695\u002Faea952","Abstract Accurate fruit quality assessment is important for selective harvesting in agricultural robotics, where multispectral sensing provides a non-destructive approach but remains sen- sitive to variability in illumination conditions. This study presents a controlled laboratory evaluation conducted in a dark box under halogen illumination with fixed sensor and sam- ple geometry, in which illumination variation was simulated by adjusting lamp intensity. Prediction models for visually assigned grape quality grades were first established under fixed illumination using an AS7265x multispectral sensor with 18 channels from 410 to 940 nm, covering the visible and part of the near-infrared range. Spectral data were prepro- cessed using standard normal variate transformation, standardization, variable importance in projection feature selection, and mixup interpolation. Three machine learning meth- ods, partial least squares regression, support vector regression, and random forest, were evaluated with nested cross-validation, out-of-fold prediction, and 2000-iteration bootstrap resampling. Support vector regression achieved the best performance, with R² = 0.9084 [95% CI: 0.851–0.945], RMSE = 0.3352, and MAE = 0.2374. To address illumination vari- ability, three compensation methods, direct standardization, extended multiplicative scatter correction, and a polynomial per-band fitting method, were implemented and compared across five lamp intensity levels from 20% to 100% of rated power. The three methods and the uncompensated condition reached similar mean accuracy across the five levels, with the exception of extended multiplicative scatter correction, which was lower at every level. Di- rect standardization reduced the spread of the five predictions obtained for the same berry by about 31% relative to no compensation, and gave the highest grade-level accuracy of the four conditions. These results provide laboratory feasibility evidence that direct standard- ization can improve the consistency of grade prediction across controlled illumination levels, and establish a basis for further validation under natural field illumination toward future robotic harvesting applications.","摘要 准确的水果品质评估对于农业机器人中的选择性采收至关重要，其中多光谱传感提供了一种非破坏性方法，但仍对光照条件的变化较为敏感。本研究提出了一种在暗箱中、卤素灯照明下、传感器与样品几何构型固定的受控实验室评估方法，通过调节灯强度来模拟光照变化。首先在固定光照条件下，使用AS7265x多光谱传感器（具有410至940 nm范围内的18个通道，覆盖可见光及部分近红外波段）建立了视觉 assigned 葡萄品质等级的预测模型。光谱数据采用标准正态变量变换、标准化、投影变量重要性特征选择以及mixup插值进行预处理。采用嵌套交叉验证、折外预测和2000次迭代自助重采样，对偏最小二乘回归、支持向量回归和随机森林三种机器学习方法进行了评估。支持向量回归取得了最佳性能，R² = 0.9084 [95% CI: 0.851–0.945]，RMSE = 0.3352，MAE = 0.2374。为解决光照变化问题，实施了直接标准化、扩展多元散射校正和逐波段多项式拟合三种补偿方法，并在额定功率20%至100%的五个灯强度水平下进行了比较。三种方法及未补偿条件在五个水平下达到了相似的平均准确率，但扩展多元散射校正除外，其在每个水平下均较低。与未补偿相比，直接标准化使同一浆果五次预测结果的离散度降低了约31%，并在四种条件中给出了最高的等级准确率。这些结果为直接标准化能够提高受控光照水平下等级预测的一致性提供了实验室可行性证据，并为未来机器人采收应用中在自然田间光照下进一步验证奠定了基础。","Engineering Research Express","2026-09-17T00:00:00Z",{"impact":19,"substance":82,"depth":83,"authority":84,"freshness":19,"relevant":20,"comment":85},20,17,13,"实验室条件下多光谱葡萄品质分级与光照补偿研究，方法扎实、结论可靠，但属实验室可行性验证，产业影响有限。",[87],{"name":79,"url":76},[89,25,27,90,91],"智慧农业","多光谱检测","水果品质分级",{"code":31,"slug":32,"name":33,"path":34,"parent_code":35},[94],{"code":95,"slug":96,"name":97,"path":98,"parent_code":99},"I3","remote-sensing","遥感与农情监测","\u002Ftopic\u002Fdigital-agriculture\u002Fremote-sensing","I",{"status":38,"locked":13,"version":39,"method":40,"evidence":101},{"candidates":102,"agricultural_evidence":45},[103,109,115,122,129],{"code":31,"score":104,"supported":45,"title":105,"abstract":107,"tags":108},15,[106],"grape",[27,106],[27],{"code":95,"score":110,"supported":45,"title":111,"abstract":113,"tags":114},11,[112],"multispectral",[112],[],{"code":50,"score":116,"supported":45,"title":117,"abstract":118,"tags":121},7,[],[119,120],"机器学习","machine learning",[25],{"code":123,"score":71,"supported":45,"title":124,"abstract":125,"tags":128},"F6",[],[126,127],"农业机器人","robotic harvesting",[],{"code":130,"score":20,"supported":13,"title":131,"abstract":132,"tags":133},"I1",[],[],[89],[135,136],"多光谱 葡萄 品质分级","AS7265x 多光谱传感器","多光谱葡萄品质分级-2874",2,"10.1088\u002F2631-8695\u002Faea952",{"doi":139,"openalex_id":141,"authors":142,"venue":79,"cited_by_count":61,"oa_url":76,"card":150,"direction":66,"ingested_from":156},"W7213509882",[143,145,147],{"name":144,"orcid":8},"Guodong Xiu",{"name":146,"orcid":8},"Sára Strakošová",{"name":148,"orcid":149},"Ikuo Mizuuchi","https:\u002F\u002Forcid.org\u002F0000-0003-4657-2613",{"tldr":151,"method":152,"finding":153,"direction":154,"opportunity":155},"在暗箱中评估多光谱葡萄品质分级预测，并比较三种光照补偿方法。","AS7265x多光谱传感器、SNV与VIP预处理、SVR\u002FRF建模、直接标准化等","SVR预测最佳（R²=0.9084）；直接标准化使同果预测离散度降低约31%。","农业遥感与作物表型","可延伸至自然田间光照下的验证，并融合机器人采摘的实时多光谱分级系统。","openalex","2026-09-18T23:30:52.790038Z",{"id":159,"title":160,"url":161,"summary":162,"summary_zh":163,"content":8,"source_name":164,"source_url":161,"published_at":165,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":166,"score_detail":167,"sources":170,"tags":172,"primary_topic":174,"secondary_topics":175,"classification":181,"search_phrases":202,"slug":205,"view_count":58,"doi":206,"paper":207,"created_at":232},1278,"ERG-Mask: An edge–region collaborative instance segmentation model for greenhouse table-grape clusters","https:\u002F\u002Fdoi.org\u002F10.1016\u002Fj.compag.2026.112369","ERG-Mask: An edge–region collaborative instance segmentation model for greenhouse table-grape clusters。Computers and Electronics in Agriculture","ERG-Mask：面向温室鲜食葡萄果串的边缘-区域协同实例分割模型。《Computers and Electronics in Agriculture》","Computers and Electronics in Agriculture","2026-08-31T00:00:00Z",62,{"impact":16,"substance":17,"depth":18,"authority":168,"freshness":138,"relevant":20,"comment":169},14,"针对温室葡萄簇的实例分割新方法，发表于核心期刊，技术有创新性，但影响范围限于细分领域，时效性较低。",[171],{"name":164,"url":161},[89,25,173,27],"设施农业",{"code":31,"slug":32,"name":33,"path":34,"parent_code":35},[176],{"code":177,"slug":178,"name":179,"path":180,"parent_code":35},"B7","protected-horticulture","设施园艺与植物工厂","\u002Ftopic\u002Fhorticulture\u002Fprotected-horticulture",{"status":38,"locked":13,"version":39,"method":40,"evidence":182},{"candidates":183,"agricultural_evidence":45},[184,188,194,198],{"code":31,"score":104,"supported":45,"title":185,"abstract":186,"tags":187},[106],[27,106],[27],{"code":177,"score":168,"supported":45,"title":189,"abstract":191,"tags":193},[190],"greenhouse",[192,190],"温室",[],{"code":130,"score":20,"supported":13,"title":195,"abstract":196,"tags":197},[],[],[89],{"code":50,"score":20,"supported":13,"title":199,"abstract":200,"tags":201},[],[],[25],[203,204],"农业人工智能 智慧农业 设施农业 葡萄","农业人工智能 智慧农业","农业人工智能智慧农业设施农业葡萄-1278","10.1016\u002Fj.compag.2026.112369",{"doi":206,"openalex_id":208,"authors":209,"venue":164,"cited_by_count":61,"oa_url":8,"card":227,"direction":66,"ingested_from":156},"W7204777248",[210,213,216,219,221,224],{"name":211,"orcid":212},"Zhengtong Ning","https:\u002F\u002Forcid.org\u002F0000-0002-9218-6425",{"name":214,"orcid":215},"Jian Li","https:\u002F\u002Forcid.org\u002F0000-0002-4187-317X",{"name":217,"orcid":218},"Haidong Li","https:\u002F\u002Forcid.org\u002F0000-0002-6634-9593",{"name":220,"orcid":8},"Liangkuan Zhu",{"name":222,"orcid":223},"Qingjie Wang","https:\u002F\u002Forcid.org\u002F0009-0001-9038-7821",{"name":225,"orcid":226},"Yi Wang","https:\u002F\u002Forcid.org\u002F0000-0002-0192-3195",{"tldr":228,"method":229,"finding":230,"direction":66,"opportunity":231},"提出ERG-Mask模型，用于温室鲜食葡萄串的实例分割，提升边缘与区域协同精度。","结合边缘与区域信息的协同实例分割模型，基于深度学习。","ERG-Mask在葡萄串分割上优于现有方法，精度更高。","可探索该模型在复杂遮挡、不同光照下的鲁棒性，或扩展至其他簇状果实（如番茄、蓝莓）的检测。","2026-09-01T23:30:01.626379Z",{"id":234,"title":235,"url":236,"summary":237,"summary_zh":8,"content":238,"source_name":239,"source_url":8,"published_at":240,"category":241,"cover_url":8,"hotness":12,"is_selected":13,"score":242,"score_detail":243,"sources":245,"tags":247,"primary_topic":251,"secondary_topics":252,"classification":253,"search_phrases":269,"slug":272,"view_count":61,"doi":8,"paper":8,"created_at":273},5516,"上海农科院葡萄团队赴银川参加第47届世界葡萄与葡萄酒大会","https:\u002F\u002Fwww.saas.sh.cn\u002Fhzjl\u002Fgjhz\u002Fcontent_57128","10月12日至16日第47届世界葡萄与葡萄酒大会在银川举办，智慧农业研究中心葡萄团队负责人奚晓军研究员作专题报告，阐明疏穗与环剥处理对果实糖、花青素及脱落酸积累的源库调控机制。团队还观摩酿酒葡萄AI综合作业机器人成果展示，与研发单位围绕葡萄智慧装备与果园智能化管理深入交流。","2026年10月12-16日，第47届世界葡萄与葡萄酒大会在银川举办，智慧农业研究中心葡萄团队负责人奚晓军研究员一行赴会交流。团队围绕“提升葡萄种植生产韧性与可持续发展能力”主题，向全球同行展示最新科研成果。\n\n会上，奚晓军作“Changes in sugar, anthocyanin and ABA accumulation in response to cluster thinning and girdling in ‘Jumeigui’grape”专题报告。报告围绕葡萄疏穗与环剥处理开展研究，阐明源库比对果实糖、花青素及脱落酸积累的调控机制，为葡萄优质栽培技术提供理论支撑，报告内容引发国内外园艺专家广泛研讨。参会期间，团队前往银川国际会展中心观摩酿酒葡萄 AI 综合作业机器人成果展示，详细了解该装备在葡萄园智能作业方面的技术特点与应用前景，并与研发单位围绕葡萄智慧装备研发、果园智能化管理等方向开展深入交流。\n\n本次参会是葡萄团队持续推进科研国际化、对标国际前沿开展研究的重要实践。","上海市农业科学院 2026-10-19","2026-10-19T00:00:00Z","报道",65,{"impact":18,"substance":83,"depth":84,"authority":84,"freshness":71,"relevant":20,"comment":244},"国际学术会议交流并涉及葡萄AI作业机器人观摩，有实质科研内容与行业参考价值，但属团队参会动态，影响面有限。",[246],{"name":239,"url":236},[89,248,27,249,250],"智能农机","栽培技术","国际交流",{"code":31,"slug":32,"name":33,"path":34,"parent_code":35},[],{"status":38,"locked":13,"version":39,"method":40,"evidence":254},{"candidates":255,"agricultural_evidence":45},[256,260,264],{"code":31,"score":16,"supported":45,"title":257,"abstract":258,"tags":259},[27],[27],[27],{"code":130,"score":44,"supported":45,"title":261,"abstract":262,"tags":263},[],[89],[89],{"code":265,"score":20,"supported":13,"title":266,"abstract":267,"tags":268},"F5",[],[],[248],[270,271],"上海农科院 葡萄 银川","世界葡萄与葡萄酒大会 奚晓军","上海农科院葡萄银川-5516","2026-10-22T00:07:04.061911Z",{"id":275,"title":276,"url":277,"summary":278,"summary_zh":279,"content":8,"source_name":280,"source_url":277,"published_at":281,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":282,"score_detail":283,"sources":286,"tags":288,"primary_topic":292,"secondary_topics":293,"classification":294,"search_phrases":321,"slug":324,"view_count":61,"doi":325,"paper":326,"created_at":347},5401,"A HIERARCHICAL DISTRIBUTED EDGE AI FRAMEWORK FOR REAL-TIME APPLE LEAF DISEASE CLASSIFICATION","https:\u002F\u002Fdoi.org\u002F10.22190\u002Ffuacr260506002g","The application of artificial intelligence (AI) in agriculture has transformed crop management, particularly in the detection of plant diseases. Apple trees, a key agricultural commodity, are vulnerable to various diseases, which can severely impact crop yields. Traditional disease detection methods, often reliant on manual inspections, are time-consuming and prone to error. In response, this study proposes a novel two-stage edge AI framework for apple leaf disease classification that integrates RGB imagery with distributed edge computing. The first stage performs binary classification to distinguish between healthy and diseased leaves, achieving an F1 score of 95.50%. This optimization is achieved by utilizing lightweight models on edge devices, thereby enhancing computational efficiency. The second stage employs a consensus-based multiclass classification approach, achieving a 96.44% F1 score in identifying specific diseases. This dual-stage system, supported by an edge AI architecture, ensures real-time disease detection while minimizing resource consumption and enhancing scalability in large-scale agricultural environments. The full system achieved a 95.50% accuracy and F1 score on the special dataset. The results demonstrate the system’s efficacy in controlled laboratory conditions, offering a promising solution for precise, real-time agricultural monitoring. This study underscores the potential of distributed edge AI systems for scalable and efficient disease detection in precision agriculture.","人工智能（AI）在农业中的应用已深刻改变了作物管理方式，尤其是在植物病害检测方面。苹果树作为重要的农业经济作物，易受多种病害侵袭，严重影响作物产量。传统的病害检测方法通常依赖人工检查，耗时且易出错。为此，本研究提出了一种新颖的两阶段边缘AI框架，用于苹果叶片病害分类，该框架将RGB图像与分布式边缘计算相结合。第一阶段执行二分类以区分健康叶片与患病叶片，F1分数达到95.50%。该优化通过在边缘设备上采用轻量级模型实现，从而提升了计算效率。第二阶段采用基于共识的多类分类方法，在识别具体病害方面F1分数达到96.44%。该双阶段系统在边缘AI架构的支持下，实现了实时病害检测，同时最大限度地降低了资源消耗，并增强了在大规模农业环境中的可扩展性。完整系统在专用数据集上达到了95.50%的准确率和F1分数。结果表明，该系统在受控实验室条件下具有有效性，为精准、实时的农业监测提供了一种有前景的解决方案。本研究凸显了分布式边缘AI系统在精准农业中实现可扩展且高效病害检测的潜力。","Facta Universitatis Series Automatic Control and Robotics","2026-10-18T00:00:00Z",76,{"impact":17,"substance":284,"depth":83,"authority":16,"freshness":19,"relevant":20,"comment":285},21,"该研究提出两级边缘AI框架，在苹果叶片病害识别上取得95%以上F1值，方法新颖、数据可靠，对精准农业实时监测有参考价值，但尚处实验室验证阶段，产业影响有限。",[287],{"name":280,"url":277},[89,25,289,290,291],"边缘计算","作物病害识别","苹果病害",{"code":31,"slug":32,"name":33,"path":34,"parent_code":35},[],{"status":38,"locked":13,"version":39,"method":40,"evidence":295},{"candidates":296,"agricultural_evidence":45},[297,303,309,315],{"code":31,"score":17,"supported":45,"title":298,"abstract":300,"tags":302},[299],"apple",[33,301,299],"苹果",[291],{"code":130,"score":116,"supported":45,"title":304,"abstract":305,"tags":308},[],[306,307],"精准农业","precision agriculture",[89],{"code":310,"score":116,"supported":45,"title":311,"abstract":312,"tags":314},"I5",[],[289,313],"edge computing",[289],{"code":50,"score":316,"supported":45,"title":317,"abstract":318,"tags":320},5,[],[319],"artificial intelligence",[290,25],[322,323],"苹果叶片病害 边缘AI","分布式边缘计算 病害分类","苹果叶片病害边缘AI-5401","10.22190\u002Ffuacr260506002g",{"doi":325,"openalex_id":327,"authors":328,"venue":280,"cited_by_count":61,"oa_url":277,"card":341,"direction":66,"ingested_from":156},"W7222964752",[329,332,335,338],{"name":330,"orcid":331},"Milan Grujev","https:\u002F\u002Forcid.org\u002F0009-0002-3211-6844",{"name":333,"orcid":334},"Dragan H. Stojanovic","https:\u002F\u002Forcid.org\u002F0000-0003-4893-3814",{"name":336,"orcid":337},"Miloš Ilić","https:\u002F\u002Forcid.org\u002F0000-0002-9696-2874",{"name":339,"orcid":340},"Aleksandar Milosavljević","https:\u002F\u002Forcid.org\u002F0000-0003-3057-7727",{"tldr":342,"method":343,"finding":344,"direction":345,"opportunity":346},"提出两级边缘AI框架，用RGB图像实时分类苹果叶片病害。","分布式边缘计算+轻量模型，先二分类再共识多分类。","两级系统准确率与F1均达95.50%，多分类F1为96.44%。","智慧农业 \u002F 农业物联网","仅在受控实验室验证，可延伸至田间复杂光照与多作物泛化及边缘能耗优化。","2026-10-20T23:30:33.830748Z",{"id":349,"title":350,"url":351,"summary":352,"summary_zh":8,"content":8,"source_name":353,"source_url":8,"published_at":354,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":355,"score_detail":356,"sources":359,"tags":361,"primary_topic":364,"secondary_topics":365,"classification":366,"search_phrases":385,"slug":388,"view_count":61,"doi":8,"paper":389,"created_at":396},5130,"DefoliationCAM:面向葡萄藤食叶严重度实时评估的边缘AI系统 [预印本]","https:\u002F\u002Farxiv.org\u002Fabs\u002F2610.20879","Fuchen Li等提出DefoliationCAM(D-CAM),一种太阳能供电的边缘AI系统,用于估算葡萄藤树冠食叶程度并通过LoRa将结果传输至Web仪表盘。流水线使用YOLOX-Nano在4K图像中定位葡萄藤主干,并使用基于MobileNet的轻量级通道注意力分类器对320×320像素的斑块进行分类。29,674个田间采集斑块被标注为背景或四种损伤等级之一;合并中度和重度类别将类别不平衡比从约40:1降至8.8:1。合成增强使macro-F1从0.768提升至0.838。分类器以更少的参数量(1.08M vs 23.52M)在macro-F1上优于ResNet-50。利用TensorRT FP16在NVIDIA Jetson Orin Nano上每张4K图像处理时间约1.5秒,完整运行周期57-60秒。两台设备为期五周的田间部署验证了自主太阳能运行、LoRa通信和远程食叶监测的可行性。","arXiv 2610.20879 2026 年 10 月 14 日","2026-10-14T00:00:00Z",75,{"impact":18,"substance":357,"depth":17,"authority":110,"freshness":19,"relevant":20,"comment":358},22,"预印本提出太阳能边缘AI系统，实现葡萄藤食叶严重度实时评估与LoRa远程监测，方法新颖、数据规模大、田间验证充分，具智慧农业应用价值。",[360],{"name":353,"url":351},[89,25,289,362,363],"葡萄藤","病害评估",{"code":31,"slug":32,"name":33,"path":34,"parent_code":35},[],{"status":38,"locked":13,"version":39,"method":40,"evidence":367},{"candidates":368,"agricultural_evidence":45},[369,373,377,381],{"code":31,"score":110,"supported":45,"title":370,"abstract":371,"tags":372},[27],[27],[],{"code":130,"score":20,"supported":13,"title":374,"abstract":375,"tags":376},[],[],[89],{"code":50,"score":20,"supported":13,"title":378,"abstract":379,"tags":380},[],[],[25],{"code":310,"score":20,"supported":13,"title":382,"abstract":383,"tags":384},[],[],[289],[386,387],"DefoliationCAM 葡萄藤 食叶","Jetson Orin Nano 边缘AI 葡萄","DefoliationCAM葡萄藤食叶-5130",{"doi":8,"openalex_id":8,"authors":390,"venue":8,"cited_by_count":61,"oa_url":8,"card":391,"direction":345,"ingested_from":68},[],{"tldr":392,"method":393,"finding":394,"direction":345,"opportunity":395},"提出太阳能边缘AI系统DefoliationCAM，实时评估葡萄藤食叶严重度并远程传输。","YOLOX-Nano定位主干，MobileNet通道注意力分类，TensorRT","轻量分类器macro-F1达0.838，优于ResNet-50，田间部署验证可行。","可探索多作物迁移、损伤等级细化及边缘模型自适应更新以提升泛化性。","2026-10-17T00:13:44.932340Z",{"id":398,"title":399,"url":400,"summary":401,"summary_zh":402,"content":8,"source_name":403,"source_url":400,"published_at":354,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":404,"score_detail":405,"sources":408,"tags":410,"primary_topic":414,"secondary_topics":418,"classification":421,"search_phrases":470,"slug":473,"view_count":20,"doi":474,"paper":475,"created_at":496},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",81,{"impact":17,"substance":357,"depth":17,"authority":168,"freshness":406,"relevant":20,"comment":407},9,"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。",[409],{"name":403,"url":400},[89,25,411,412,413],"产量预测","多模态融合","遥感",{"code":50,"slug":415,"name":416,"path":417,"parent_code":99},"agricultural-ai","农业AI与大模型","\u002Ftopic\u002Fdigital-agriculture\u002Fagricultural-ai",[419,420],{"code":31,"slug":32,"name":33,"path":34,"parent_code":35},{"code":95,"slug":96,"name":97,"path":98,"parent_code":99},{"status":38,"locked":13,"version":39,"method":40,"evidence":422},{"candidates":423,"agricultural_evidence":45},[424,430,436,441,445,452,458,464],{"code":50,"score":110,"supported":45,"title":425,"abstract":426,"tags":429},[],[427,119,120,428],"深度学习","deep learning",[25,412],{"code":31,"score":406,"supported":45,"title":431,"abstract":432,"tags":435},[],[301,433,434,299],"芒果","orchard",[],{"code":95,"score":19,"supported":45,"title":437,"abstract":438,"tags":440},[],[439,112],"无人机多光谱",[411,413],{"code":130,"score":116,"supported":45,"title":442,"abstract":443,"tags":444},[],[306,307],[89],{"code":446,"score":71,"supported":45,"title":447,"abstract":448,"tags":451},"H2",[],[449,450],"土壤","soil",[],{"code":310,"score":71,"supported":45,"title":453,"abstract":454,"tags":457},[],[455,456],"物联网","IoT",[],{"code":459,"score":58,"supported":45,"title":460,"abstract":461,"tags":463},"H6",[],[462],"农业生态",[],{"code":465,"score":58,"supported":45,"title":466,"abstract":467,"tags":469},"H7",[],[468],"气候适应",[],[471,472],"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":474,"openalex_id":476,"authors":477,"venue":403,"cited_by_count":61,"oa_url":400,"card":491,"direction":345,"ingested_from":156},"W7221795850",[478,481,483,485,487,489],{"name":479,"orcid":480},"Shabir Ahmad Mir","https:\u002F\u002Forcid.org\u002F0000-0003-2311-218X",{"name":482,"orcid":8},"Umarfarooque Momin",{"name":484,"orcid":8},"Sudeep Kumar E",{"name":486,"orcid":8},"Itfaq Ahmad Mir",{"name":488,"orcid":8},"Ab Rouf Khan",{"name":490,"orcid":8},"Ghulam Mohammad Mir",{"tldr":492,"method":493,"finding":494,"direction":66,"opportunity":495},"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。","2026-10-15T23:30:08.750140Z"]