[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-2289":3,"related-2289":100},{"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":38,"classification":39,"search_phrases":65,"slug":68,"view_count":22,"doi":69,"paper":70,"created_at":99},2289,"High-throughput phenotyping platform for facility crops based on optical sensing technology: A review","https:\u002F\u002Fdoi.org\u002F10.1016\u002Fj.atech.2026.102545","High-throughput acquisition of crop phenotypic information is one of the key technologies for achieving intelligent facility agriculture and precision breeding. Traditional phenotypic data collection methods suffer from low efficiency and strong subjectivity, making it difficult to achieve multi-scale continuous monitoring and meet the demands of modern research and production. This paper systematically reviews the technological framework and development trajectory of optical sensing technology-driven phenotypic platforms for facility crops. First, starting from optical sensing technologies, a comparative analysis highlights the advantages and limitations of RGB, multi-\u002Fhyperspectral, thermal infrared, and LiDAR sensors in phenotypic perception. Second, the characteristics and applicable scenarios of stationary, rail-mounted, mobile robot, and unmanned aerial vehicle (UAV) platform architectures are summarized. Furthermore, the evolution of phenotypic data processing methods is examined, focusing on the shift from traditional feature engineering to deep learning-driven approaches. Finally, key challenges such as multimodal data fusion, system cost, and real-time performance are discussed, along with the future direction of phenotypic platforms toward intelligent closed-loop decision-making systems. This article systematically reviews the facility agriculture phenotyping platforms driven by optical sensing technology, and also incorporates representative research progress in field phenotyping studies. These advances provide transferable sensing technologies, methodological frameworks, and platform design concepts that can facilitate the development of phenotyping platforms for controlled-environment agriculture.","高通量获取作物表型信息是实现智能设施农业和精准育种的关键技术之一。传统表型数据采集方式效率低、主观性强，难以实现多尺度连续监测，无法满足现代研究与生产需求。本文系统综述了光学传感技术驱动的设施作物表型平台的技术框架与发展脉络。首先，从光学传感技术出发，对比分析了RGB、多光谱\u002F高光谱、热红外和激光雷达传感器在表型感知中的优势与局限。其次，总结了固定式、轨道式、移动机器人和无人机平台架构的特点与适用场景。进而，梳理了表型数据处理方法的演进，重点分析了从传统特征工程到深度学习驱动方法的转变。最后，讨论了多模态数据融合、系统成本和实时性等关键挑战，并展望了表型平台向智能闭环决策系统发展的未来方向。本文系统综述了光学传感技术驱动的设施农业表型平台，同时纳入了田间表型研究中的代表性进展。这些进展提供了可迁移的传感技术、方法框架和平台设计理念，有助于推动受控环境农业表型平台的发展。",null,"Smart Agricultural Technology","2026-09-10T00:00:00Z","论文",10,false,74,{"impact":17,"substance":18,"depth":19,"authority":20,"freshness":21,"relevant":22,"comment":23},18,20,17,13,6,1,"系统综述光学传感驱动的设施作物高通量表型平台，涵盖传感器对比、平台架构与深度学习数据处理演进，方法框架清晰、可迁移性强，对智慧设施农业与精准育种有参考价值。",[25],{"name":10,"url":6},[27,28,29,30,31],"智慧农业","农业人工智能","设施农业","高通量表型","光学传感",{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},"I2","agricultural-ai","农业AI与大模型","\u002Ftopic\u002Fdigital-agriculture\u002Fagricultural-ai","I",[],{"status":40,"locked":14,"version":41,"method":42,"evidence":43},"classified","2026-10-22.2","rules",{"candidates":44,"agricultural_evidence":47},[45,53,60],{"code":33,"score":46,"supported":47,"title":48,"abstract":49,"tags":52},7,true,[],[50,51],"深度学习","deep learning",[28],{"code":54,"score":21,"supported":47,"title":55,"abstract":56,"tags":59},"I3",[],[57,58],"作物表型","hyperspectral",[],{"code":61,"score":22,"supported":14,"title":62,"abstract":63,"tags":64},"I1",[],[],[27],[66,67],"农业人工智能 高通量表型 光学传感 智慧农业","农业人工智能 高通量表型","农业人工智能高通量表型光学传感智慧农业-2289","10.1016\u002Fj.atech.2026.102545",{"doi":69,"openalex_id":71,"authors":72,"venue":10,"cited_by_count":90,"oa_url":91,"card":92,"direction":96,"ingested_from":98},"W7212130452",[73,75,77,79,81,83,86,88],{"name":74,"orcid":9},"Xiaodong Zhang",{"name":76,"orcid":9},"Zhaowei Li",{"name":78,"orcid":9},"Yi Zhang",{"name":80,"orcid":9},"Chuandong Guo",{"name":82,"orcid":9},"Zonghua Leng",{"name":84,"orcid":85},"Xiangyu Han","https:\u002F\u002Forcid.org\u002F0000-0003-0412-9859",{"name":87,"orcid":9},"Hanping Mao",{"name":89,"orcid":9},"Yixue Zhang",0,"https:\u002F\u002Fwww.sciencedirect.com\u002Fscience\u002Farticle\u002Fpii\u002FS2772375526007707\u002Fpdf",{"tldr":93,"method":94,"finding":95,"direction":96,"opportunity":97},"综述光学传感驱动的设施作物高通量表型平台，涵盖传感器、平台架构与数据处理。","综述RGB、多\u002F高光谱、热红外、LiDAR及固定\u002F轨道\u002F机器人\u002F无人机平台与深度","表型平台正从传统特征工程转向深度学习，并迈向智能闭环决策系统。","农业遥感与作物表型","多模态数据融合、低成本实时表型平台及闭环决策在设施农业中仍待突破。","openalex","2026-09-13T23:30:04.259059Z",{"total":21,"page":22,"page_size":21,"items":101},[102,172,216,283,357,419],{"id":103,"title":104,"url":105,"summary":106,"summary_zh":107,"content":9,"source_name":108,"source_url":105,"published_at":109,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":110,"score_detail":111,"sources":116,"tags":118,"primary_topic":121,"secondary_topics":122,"classification":123,"search_phrases":144,"slug":147,"view_count":148,"doi":149,"paper":150,"created_at":171},4108,"Visión artificial para la detección de plagas en cultivos de tomate: revisión sistemática de la literatura","https:\u002F\u002Fdoi.org\u002F10.36825\u002Friti.14.34.010","Phytosanitary monitoring is an essential activity for reducing production losses caused by pests in tomato crops grown under protected agriculture conditions. In recent years, computer vision and deep learning methods have gained relevance as tools capable of automating agricultural image detection and analysis processes. Despite these advances, there is still a need to determine which techniques have been used, under what conditions they have been evaluated, and what performance they have reported, particularly in protected agriculture and small-scale production contexts. This study aims to examine research published on these technologies and assess their potential incorporation into participatory agricultural monitoring schemes. To this end, a systematic review of scientific literature published between 2020 and 2026 was conducted following the methodological guidelines proposed by Kitchenham. The review was developed through the stages of planning, study search and selection, and information synthesis, considering previously established inclusion and exclusion criteria. As a result, fifty studies related to image processing, artificial intelligence, and phytosanitary monitoring were selected and analyzed. Among the main findings of the review, several studies reported high-performance levels for deep learning models in tasks involving the automatic identification of symptoms associated with pests and other crop-related conditions. The synthesis identified the main techniques, data, application contexts, and reported performance levels, providing reference elements for assessing their potential incorporation into participatory agricultural monitoring schemes.","植物检疫监测是减少保护地番茄种植中由有害生物造成的产量损失的一项重要活动。近年来，计算机视觉和深度学习方法作为能够自动化农业图像检测与分析流程的工具而日益受到重视。尽管取得了这些进展，但仍需明确已采用了哪些技术、这些技术在何种条件下进行了评估，以及它们报告了怎样的性能，尤其是在保护地农业和小规模生产情境下。本研究旨在考察已发表的关于这些技术的研究，并评估其纳入参与式农业监测方案的潜力。为此，按照Kitchenham提出的方法学指南，对2020年至2026年间发表的科学文献进行了系统综述。该综述通过规划、研究检索与筛选以及信息综合等阶段展开，并依据事先确立的纳入与排除标准。结果共筛选并分析了50项与图像处理、人工智能和植物检疫监测相关的研究。在综述的主要发现中，多项研究报告了深度学习模型在自动识别与有害生物相关的症状及其他作物相关状况的任务中具有高性能水平。该综合识别了主要技术、数据、应用情境及所报告的性能水平，为评估其纳入参与式农业监测方案的潜力提供了参考要素。","Revista de Investigación en Tecnologías de la Información","2026-10-02T00:00:00Z",63,{"impact":112,"substance":17,"depth":113,"authority":13,"freshness":114,"relevant":22,"comment":115},12,15,8,"系统综述梳理2020-2026年计算机视觉与深度学习在番茄病虫害检测中的应用，方法规范、结论有参考价值，但属细分领域综述，公共影响有限。",[117],{"name":108,"url":105},[27,28,29,119,120],"番茄种植","病虫害监测",{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},[],{"status":40,"locked":14,"version":41,"method":42,"evidence":124},{"candidates":125,"agricultural_evidence":47},[126,133,140],{"code":33,"score":13,"supported":47,"title":127,"abstract":128,"tags":132},[],[50,129,51,130,131],"计算机视觉","artificial intelligence","computer vision",[28],{"code":134,"score":46,"supported":47,"title":135,"abstract":136,"tags":139},"B1",[],[137,138],"番茄","tomato",[119],{"code":61,"score":22,"supported":14,"title":141,"abstract":142,"tags":143},[],[],[27],[145,146],"计算机视觉 番茄 病虫害检测","深度学习 设施农业 植保监测","计算机视觉番茄病虫害检测-4108",3,"10.36825\u002Friti.14.34.010",{"doi":149,"openalex_id":151,"authors":152,"venue":108,"cited_by_count":90,"oa_url":105,"card":165,"direction":169,"ingested_from":98},"W7216085028",[153,155,158,161,163],{"name":154,"orcid":9},"Jesús Alberto Alarcón Luna",{"name":156,"orcid":157},"María Salomé Alejandre Apolinar","https:\u002F\u002Forcid.org\u002F0009-0008-0232-8217",{"name":159,"orcid":160},"Hugo Amores Pérez","https:\u002F\u002Forcid.org\u002F0009-0009-0260-4872",{"name":162,"orcid":9},"Jacel Adame García",{"name":164,"orcid":9},"María Angélica Cerdán",{"tldr":166,"method":167,"finding":168,"direction":169,"opportunity":170},"系统综述2020-2026年计算机视觉与深度学习在番茄设施栽培虫害检测中的研究。","遵循Kitchenham指南系统检索筛选50篇文献，归纳技术、数据与性能。","深度学习模型在虫害症状自动识别任务中普遍报告高性能，但多限于特定条件。","农业人工智能与决策模型","面向设施农业与小规模生产的轻量化模型及参与式监测集成仍属研究空白。","2026-10-03T23:30:53.163959Z",{"id":173,"title":174,"url":175,"summary":176,"summary_zh":9,"content":177,"source_name":178,"source_url":9,"published_at":179,"category":180,"cover_url":9,"hotness":13,"is_selected":14,"score":181,"score_detail":182,"sources":186,"tags":188,"primary_topic":191,"secondary_topics":192,"classification":193,"search_phrases":211,"slug":214,"view_count":148,"doi":9,"paper":9,"created_at":215},2729,"WAFI2026世界农业科技创新大会在京举行——辽宁省农科院等亮相","https:\u002F\u002Fwww.toutiao.com\u002Farticle\u002F7686105927610630691\u002F","2026世界农业科技创新大会(WAFI2026)9月16—19日在北京平谷举行。中国农业大学校长孙其信指出,人工智能已经从过去的示范,变成了部分大型新型经营主体的主要生产方式。中国农业大学王耀君副教授介绍神农大模型3.0已在非洲落地。北京数智京园智慧设施管控技术体系亮相。","潮新闻 记者 沈爱群 侴雪妍\n\n2026世界农业科技创新大会，正在北京举行。\n\n9月16日上午，在大会举行的“人工智能与农业论坛”上，潮新闻记者捕捉到了一个话题：人工智能如何造福农民？\n\n这个话题，由中国农业大学全球食物经济与政策研究院院长樊胜根教授在论坛致辞中提出。\n\n会上，与会国内外专家学者、业界代表见仁见智给出了答案：通过农业人工智能，可以让农民种得好、种得起、种得稳、种得赚。其中，深谙东方智慧的“中国方案”得到了与会嘉宾的点赞与关注。\n\n世界农业科技创新大会（英文缩写WAFI），是以“创新农业 共享未来”为宗旨的世界农业盛会，致力于打造农业“达沃斯”。自2023年成功举办以来，WAFI赢得了国内外同行的高度认可，被誉为世界三大农业盛会之一。\n\n![Image 1](https:\u002F\u002Fp3-sign.toutiaoimg.com\u002Ftos-cn-i-axegupay5k\u002F59a79b9b009545259d6d2d50486c2167~tplv-tt-origin-web:gif.jpeg?_iz=58558&from=article.pc_detail&lk3s=953192f4&x-expires=1790208715&x-signature=Mbp32uTVtyQxX7rWipDc9RKtXnU%3D)\n活动现场（记者 沈爱群 摄）\n\n在这个世界级农业盛会上，与会嘉宾为何特别关注“人工智能与农民”？\n\n答案，可以从人工智能时代全球农民、小农户面临的挑战找到。\n\n农业人工智能论坛上，中国农业大学校长、中国工程院院士陈卫就谈到：当前全球农食系统面临前所未有的挑战，气候变化加剧了农业生产的风险，土地、水资源和生物多样性承受着越来越大的压力。“我们必须生产更多更有营养的食物，降低农业对环境的影响，为农村地区创造更好的发展机会。然而这些挑战在不同地区的分布并不均衡，一端是资本和技术高度密集的现代化农场，另一边是数以亿计、以有限的资源支撑全球重要粮食供给的农民与小农户。”\n\n挑战面前，人工智能为农业转型注入了新动能。正如与会嘉宾在本次论坛上提及，人工智能正在推动智能育种、精准种植和农业生产全产业链系统发生新的变化，以卫星遥感数据、气象预警数据、土壤变化数据、种子种植数据以及营销数据分析等，帮助农民实现种植方案自动生成、无人机出苗率检测、卫星遥感旱涝、摄像头自动巡田、智能拼车等农业生产及农事经营。\n\n推动这些新变化的“中国方案”中，有着力农业教育的中国高校，有从事农业生产的中国农业企业，也有站在消费端的中国城市设施农业。\n\n先看中国高校。会上，中国农业大学信息与电子工程学院副教授王耀君和中国农业大学全球食物经济与政策研究院教授张玉梅，分别在主旨演讲中提到了“神农大模型”以及“农业食物经济与政策AI模型”。\n\n神农大模型，去年就已做到了3.0版。作为国内首个实现农业系统智能的大模型，神农1.0版于2023年问世，可以实现农业专业知识的精准查询与问答。2024年迭代的神农2.0版，拓展了技术边界，可以整合文本、图像等多模态数据进行分析决策。2025年全球首发的神农3.0版，打破了农业学科壁垒，让AI成为汇聚农业智慧的载体。\n\n神农大模型3.0是“小麦育种智能助手”。它融合了国家级种质资源与专家知识，可以赋能育种决策，实现从“田间试错”到“精准育种”。\n\n神农3.0还是个农业病虫害智能体。可以识别70类、600余种病虫害，实现用一部手机就能获得如同多个专业植保专家的指导。\n\n值得一提的是，神农3.0还可以让全球农业科技人员和从业者自主构建，以最低成本推动AI应用，让农业AI在科研院所和田间地头普惠落地。“经过过去一年的推广，神农大模型已经跨越千山万水，在非洲落地了。”神农大模型团队负责人王耀君说。\n\n农业食物经济与政策AI模型，不仅面向政府、科研机构和企业，也面向广大农业生产者，旨在将数据、经济模型与人工智能相结合，为农业市场研判和科学决策提供支持。为此，张玉梅教授提醒：“对于农民和小农户来说，这个模型提供的国内外农产品价格监测与异常预警、农业生产成本收益分析、膳食营养评价和国际市场动态分析，可以帮助他们及时了解市场变化 、比较政策方案、评估生产经营收益与营养状况。”\n\n其次，看看农业企业端。会上，北大荒信息有限公司总经理任荣荣向大家介绍了“未来农场”这个各项农艺技术集成平台。以深耕智慧农业培育发展新动能为己任的北大荒信息有限公司，既自主研发了智能装备管理平台覆盖111个农场、接入8.4万台智能装备；也让“未来农场”为60万种植户提供产前、产中、产后服务，实现农业资金交易1000亿元。\n\n再看城市设施农业。会上，北京市数字农业农村促进中心副主任、正高级农艺师芦天罡，向大家展示了北京市“数智京园”智慧设施管控技术体系和连栋温室“赛马制”中试熟化场景。\n\n据芦天罡介绍，北京市目前正通过AI+城市设施农业，实现了连栋温室的国产化技术攻关和日光温室的数智场景改造，快速推动农业产业智能化。“从系统到装备到模型，人工智能可以帮助农户解放劳动力、提高精细化生产水平，还可以调节生产周期，助推农户增收。”芦天罡表示。\n\n“转载请注明出处”","今日头条 2026年9月16日","2026-09-16T00:00:00Z","报道",78,{"impact":183,"substance":18,"depth":19,"authority":114,"freshness":184,"relevant":22,"comment":185},24,9,"世界级农业盛会现场报道，汇聚神农大模型3.0、未来农场、数智京园等多方实质进展，信息增量足，值得进入每日精选。",[187],{"name":178,"url":175},[27,28,29,189,190],"神农大模型","智能育种",{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},[],{"status":40,"locked":14,"version":41,"method":42,"evidence":194},{"candidates":195,"agricultural_evidence":47},[196,202,207],{"code":33,"score":197,"supported":47,"title":198,"abstract":199,"tags":201},5,[],[200],"大模型",[28,189],{"code":203,"score":22,"supported":14,"title":204,"abstract":205,"tags":206},"F2",[],[],[190],{"code":61,"score":22,"supported":14,"title":208,"abstract":209,"tags":210},[],[],[27],[212,213],"农业人工智能 神农大模型 智慧农业 智能育种","农业人工智能 神农大模型","农业人工智能神农大模型智慧农业智能育种-2729","2026-09-17T00:04:39.330857Z",{"id":217,"title":218,"url":219,"summary":220,"summary_zh":221,"content":9,"source_name":222,"source_url":219,"published_at":179,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":15,"score_detail":223,"sources":225,"tags":227,"primary_topic":230,"secondary_topics":236,"classification":238,"search_phrases":267,"slug":270,"view_count":148,"doi":271,"paper":272,"created_at":282},2682,"Machine Learning-Based Decision Support System for Greenhouse Crop Management Under Mite Infestation Conditions","https:\u002F\u002Fdoi.org\u002F10.36099\u002Fjess.v1i3.001","This paper presents an integrated machine learning-based Decision Support System (DSS) for greenhouse crop management under mite infestation conditions, specifically designed for Sri Lankan agricultural contexts. The research addresses critical challenges faced by greenhouse farmers regarding pest management and crop productivity optimization through a comprehensive system combining NetLogo simulation for synthetic data generation, ensemble machine learning models, and a web-based interface. Field research with Sri Lankan greenhouse farmers revealed that mite infestations cause up to 40% crop losses, driving panic-induced pesticide overuse and knowledge gaps in pest management timing. The system integrates environmental monitoring, pest prediction, and crop yield forecasting to provide actionable recommendations for farmers. The mite infestation prediction model achieved 86% accuracy, while the system successfully addresses data scarcity challenges through agent-based modeling. The DSS demonstrates potential for transforming reactive farming practices into predictive, data-driven approaches while accommodating the technological constraints of developing agricultural contexts.","本文提出了一种基于机器学习的集成决策支持系统（DSS），用于螨虫侵染条件下的温室作物管理，专为斯里兰卡农业情境设计。该研究针对温室农户在害虫管理和作物生产力优化方面面临的关键挑战，通过一个综合系统加以解决，该系统结合了用于合成数据生成的NetLogo仿真、集成机器学习模型以及基于网络的界面。针对斯里兰卡温室农户的实地研究表明，螨虫侵染可导致高达40%的作物损失，进而引发恐慌性农药过度使用以及害虫管理时机方面的知识缺口。该系统整合了环境监测、害虫预测和作物产量预测，为农户提供可操作的推荐建议。螨虫侵染预测模型达到了86%的准确率，同时该系统通过基于智能体的建模成功应对了数据稀缺的挑战。该决策支持系统展现出将被动应对式耕作实践转变为预测性、数据驱动方法的潜力，同时兼顾了发展中农业情境的技术约束。","Journal of Environmental and Sustainability Science",{"impact":113,"substance":18,"depth":19,"authority":112,"freshness":13,"relevant":22,"comment":224},"将机器学习与智能体仿真结合用于温室螨害预测与决策支持，方法新颖、数据翔实，对设施农业植保信息化有参考价值。",[226],{"name":222,"url":219},[27,28,29,228,229],"决策支持系统","病虫害预警",{"code":231,"slug":232,"name":233,"path":234,"parent_code":235},"B7","protected-horticulture","设施园艺与植物工厂","\u002Ftopic\u002Fhorticulture\u002Fprotected-horticulture","B",[237],{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},{"status":40,"locked":14,"version":41,"method":42,"evidence":239},{"candidates":240,"agricultural_evidence":47},[241,249,256,263],{"code":33,"score":242,"supported":47,"title":243,"abstract":246,"tags":248},27,[244,245],"machine learning","decision support",[247,244,245],"机器学习",[28,228],{"code":231,"score":250,"supported":47,"title":251,"abstract":253,"tags":255},14,[252],"greenhouse",[254,252],"温室",[],{"code":257,"score":21,"supported":47,"title":258,"abstract":259,"tags":262},"F4",[],[260,261],"农药","pesticide",[],{"code":61,"score":22,"supported":14,"title":264,"abstract":265,"tags":266},[],[],[27],[268,269],"农业人工智能 决策支持系统 病虫害预警 智慧农业","农业人工智能 决策支持系统","农业人工智能决策支持系统病虫害预警智慧农业-2682","10.36099\u002Fjess.v1i3.001",{"doi":271,"openalex_id":273,"authors":274,"venue":222,"cited_by_count":90,"oa_url":219,"card":277,"direction":169,"ingested_from":98},"W7213235680",[275],{"name":276,"orcid":9},"S. Nasiketha",{"tldr":278,"method":279,"finding":280,"direction":169,"opportunity":281},"为斯里兰卡温室农户开发基于机器学习的决策支持系统，预测螨害并优化作物管理。","NetLogo仿真生成合成数据，集成机器学习模型与网页界面。","螨害预测准确率达86%，可缓解数据稀缺并减少农药滥用。","可探索小样本下合成数据与迁移学习结合，提升发展中国家温室病虫害预测泛化能力。","2026-09-16T23:30:51.573509Z",{"id":284,"title":285,"url":286,"summary":287,"summary_zh":288,"content":9,"source_name":289,"source_url":286,"published_at":290,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":291,"score_detail":292,"sources":294,"tags":296,"primary_topic":299,"secondary_topics":304,"classification":306,"search_phrases":324,"slug":326,"view_count":148,"doi":327,"paper":328,"created_at":356},2523,"Deep learning-based detection and counting of wheat seeds: Comparative benchmarking of YOLO models","https:\u002F\u002Fdoi.org\u002F10.56612\u002Fijaaeb.v6i1.249","Automated wheat-seed detection and counting play important roles in high-throughput plant phenotyping, seed characterization, and agricultural research. Conventional manual counting is labor-intensive, time-consuming, and susceptible to human error when processing large numbers of seed samples. Recent advances in deep learning and object detection provide opportunities to automate these tasks using conventional RGB images. This study focused on a deep learning-based framework for wheat-seed detection and detection-based counting using a custom red-green-blue (RGB) image dataset. A dataset comprising 832 RGB images containing 3,649 manually annotated wheat-seed instances was developed, with images representing one to ten detached wheat seeds per image. All seed instances were manually annotated using bounding boxes and formulated as a single-class object detection problem. Two lightweight object detection models, YOLOv8n and YOLO11n, were trained and evaluated under identical experimental conditions. The model performance was assessed using precision, recall, mean Average Precision at an Intersection over Union (IoU) threshold of 0.5 (mAP@0.5), mean Average Precision averaged across IoU thresholds from 0.5 to 0.95 (mAP@0.5:0.95), training loss curves, confidence-based performance curves, confusion matrices, and qualitative detection outputs. Both models achieved excellent detection performance on the custom wheat seed dataset. YOLOv8n achieved a precision of 0.9947, recall of 0.9963, mAP@0.5 of 0.9940, and mAP@0.5:0.95 of 0.5284. YOLO11n produced slightly higher performance, achieving a precision of 0.9988, recall of 0.9984, mAP@0.5 of 0.9950, and mAP@0.5:0.95 of 0.5385. Overall, the performance of the two models was similar at mAP@0.5, but at mAP@0.5:0.95 stricter criterion, YOLO11n consistently demonstrated the strongest overall performance. The findings show that lightweight YOLO models combined with RGB imaging provide an effective and easy way for automated localization and counting of wheat seeds. ​The framework provides a manually annotated RGB wheat-seed dataset and a reproducible benchmark to compare lightweight YOLO models for detection. This study provides a practical foundation for future research on automated seed phenotyping, with future work focusing on external validation using more diverse datasets, multi-class seed-quality assessment, and quantitative evaluation of counting performance.","自动化小麦种子检测与计数在高通量植物表型分析、种子表征和农业研究中发挥着重要作用。传统的人工计数在处理大量种子样本时劳动强度大、耗时长且易受人为误差影响。近年来深度学习和目标检测的进展为利用常规RGB图像实现这些任务的自动化提供了机遇。本研究聚焦于基于深度学习的框架，利用自定义红绿蓝（RGB）图像数据集进行小麦种子检测及基于检测的计数。构建了一个包含832张RGB图像、3，649个手动标注小麦种子实例的数据集，图像中每张包含一到十粒脱离的小麦种子。所有种子实例均使用边界框进行手动标注，并形式化为单类目标检测问题。在相同实验条件下训练和评估了两种轻量级目标检测模型YOLOv8n和YOLO11n。采用精确率、召回率、交并比（IoU）阈值为0.5时的平均精度均值（mAP@0.5）、IoU阈值从0.5到0.95的平均精度均值（mAP@0.5:0.95）、训练损失曲线、基于置信度的性能曲线、混淆矩阵以及定性检测输出对模型性能进行了评估。两种模型在自定义小麦种子数据集上均取得了优异的检测性能。YOLOv8n的精确率为0.9947，召回率为0.9963，mAP@0.5为0.9940，mAP@0.5:0.95为0.5284。YOLO11n的性能略高，精确率为0.9988，召回率为0.9984，mAP@0.5为0.9950，mAP@0.5:0.95为0.5385。总体而言，两种模型在mAP@0.5上的性能相似，但在mAP@0.5:0.95这一更严格的指标下，YOLO11n始终展现出最强的整体性能。研究结果表明，轻量级YOLO模型结合RGB成像为小麦种子的自动化定位和计数提供了一种有效且简便的方法。该框架提供了一个手动标注的RGB小麦种子数据集和一个可重复的基准，用于比较轻量级YOLO模型的检测性能。本研究为未来自动化种子表型分析研究提供了实用基础，未来工作将侧重于使用更多样化数据集进行外部验证、多类种子质量评估以及计数性能的定量评价。","International Journal of Applied and Experimental Biology","2026-09-14T00:00:00Z",73,{"impact":113,"substance":18,"depth":19,"authority":20,"freshness":114,"relevant":22,"comment":293},"基于自建RGB数据集对YOLOv8n与YOLO11n进行小麦种子检测计数对比，方法清晰、指标完整，属细分领域可复现基准研究，对自动化种子表型有实用参考价值。",[295],{"name":289,"url":286},[27,28,297,30,298],"小麦","种子检测",{"code":300,"slug":301,"name":297,"path":302,"parent_code":303},"A2","wheat","\u002Ftopic\u002Ffield-crops\u002Fwheat","A",[305],{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},{"status":40,"locked":14,"version":41,"method":42,"evidence":307},{"candidates":308,"agricultural_evidence":47},[309,316,320],{"code":33,"score":310,"supported":47,"title":311,"abstract":313,"tags":315},26,[51,312],"YOLO",[314,50,51,312],"目标检测",[28],{"code":300,"score":113,"supported":47,"title":317,"abstract":318,"tags":319},[301],[297,301],[297],{"code":61,"score":22,"supported":14,"title":321,"abstract":322,"tags":323},[],[],[27],[325,67],"农业人工智能 高通量表型 智慧农业 种子检测","农业人工智能高通量表型智慧农业种子检测-2523","10.56612\u002Fijaaeb.v6i1.249",{"doi":327,"openalex_id":329,"authors":330,"venue":289,"cited_by_count":90,"oa_url":286,"card":351,"direction":96,"ingested_from":98},"W7212562005",[331,334,336,339,341,343,345,347,349],{"name":332,"orcid":333},"Faisal Shahzad","https:\u002F\u002Forcid.org\u002F0009-0003-8413-2456",{"name":335,"orcid":9},"Hafiza Ayesha Arshad",{"name":337,"orcid":338},"Habib‐ur‐Rehman Athar","https:\u002F\u002Forcid.org\u002F0000-0002-8733-3865",{"name":340,"orcid":9},"Israr Hanif",{"name":342,"orcid":9},"Iqra Shokat",{"name":344,"orcid":9},"Ayesha Maryam",{"name":346,"orcid":9},"Laiba Urooj",{"name":348,"orcid":9},"Jaweria Maqbool",{"name":350,"orcid":9},"Aleena Akram",{"tldr":352,"method":353,"finding":354,"direction":96,"opportunity":355},"用自建RGB小麦种子数据集对比YOLOv8n与YOLO11n的检测计数性能。","832张RGB图像、3649个标注框，训练YOLOv8n与YOLO11n并多指标","两模型mAP@0.5均超0.994，YOLO11n在严格指标mAP@0.5:0.95上更优。","可扩展多品种、多类别种子质量评估，并引入更复杂背景与计数精度量化验证。","2026-09-15T23:30:17.202179Z",{"id":358,"title":359,"url":360,"summary":361,"summary_zh":362,"content":9,"source_name":363,"source_url":360,"published_at":364,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":365,"score_detail":366,"sources":369,"tags":371,"primary_topic":375,"secondary_topics":381,"classification":383,"search_phrases":399,"slug":402,"view_count":90,"doi":403,"paper":404,"created_at":418},5592,"Evaluation of Concept-Rank Collapse Audits for Computer Vision Agriculture Disease Classifiers Under Low-Label Regimes: An Interpretability Capacity Bound for Smallholder Fungicide Selectivity","https:\u002F\u002Fdoi.org\u002F10.5281\u002Fzenodo.23545193","Concept-based interpretability methods for plant-disease classifiers are typically evaluated by whether their concept activations correlate with human-annotated labels, but this criterion is silent on a prior question: whether the classifier's internal representation retains enough concept-rank structure for any audit to be informative. the present analysis formalises that prior question as an interpretability capacity bound. A classifier is modelled as a concept-rank operator whose singular values govern how much concept-level signal survives the projection from pixel space to class logits, and a collapse audit is defined as the ratio of retained concept-rank energy to total concept energy. Under a low-label regime, where concept annotations are scarce and the audit must be estimated from few labelled examples, the bound predicts that audit reliability degrades as a function of the concept-rank spectrum and the label budget. The analysis is a theoretical and methodological contribution: it derives the bound, states the conditions under which a collapse audit is identifiable, and characterises the operating characteristics of the audit as a decision procedure under a declared variance structure (between-cluster SD = 6.0, residual SD = 4.5, intraclass correlation = 0.64). Power was estimated across 2000 design cells for cluster counts of 8, 16, 24 and 32 and effect sizes of 3, 5 and 8 percentage points; 52 clusters were required for 5 percentage points at 80% power. The bound reframes concept audits as capacity tests rather than fidelity tests, with consequences for how smallholder fungicide-selectivity decisions should be supported by interpretable disease classifiers.","面向植物病害分类器的基于概念的可解释性方法，通常以其概念激活是否与人工标注标签相关作为评价标准，但该标准对一个更先前的问题保持沉默：分类器的内部表征是否保留了足够的概念秩结构，从而使任何审计具有信息量。本分析将该先前问题形式化为一个可解释性容量上界。分类器被建模为概念秩算子，其奇异值支配着从像素空间到类别logits的投影中概念级信号得以保留的程度，并将坍缩审计定义为保留的概念秩能量与总概念能量之比。在低标签情形下，概念标注稀缺且审计必须由少量标注样本估计，该上界预测审计可靠性会随概念秩谱和标签预算的函数而退化。本分析是一项理论和方法论贡献：推导出该上界，陈述坍缩审计可识别的条件，并在声明的方差结构（簇间SD = 6.0，残差SD = 4.5，组内相关系数 = 0.64）下刻画审计作为决策程序的操作特征。在2000个设计单元中，针对簇数为8、16、24和32以及效应量为3、5和8个百分点的情形估计了功效；在80%功效下达到5个百分点需要52个簇。该上界将概念审计重新框定为容量检验而非保真度检验，并对可解释病害分类器应如何支持小农杀菌剂选择性决策具有启示意义。","Zenodo (CERN European Organization for Nuclear Research)","2026-10-21T00:00:00Z",69,{"impact":114,"substance":367,"depth":17,"authority":112,"freshness":13,"relevant":22,"comment":368},21,"理论方法论文，提出可解释性容量上界并给出审计功效分析，对智慧农业病害识别可信度有参考价值，但属小众方法学研究、缺乏田间实证，暂不宜进入每日精选。",[370],{"name":363,"url":360},[27,28,372,373,374],"小农户","植物病害识别","可解释性",{"code":376,"slug":377,"name":378,"path":379,"parent_code":380},"G2","farm-services","社会化服务与规模经营","\u002Ftopic\u002Frural-services\u002Ffarm-services","G",[382],{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},{"status":40,"locked":14,"version":41,"method":42,"evidence":384},{"candidates":385,"agricultural_evidence":47},[386,391,395],{"code":376,"score":112,"supported":47,"title":387,"abstract":389,"tags":390},[388],"smallholder",[388],[372],{"code":33,"score":13,"supported":47,"title":392,"abstract":393,"tags":394},[131],[],[28,373],{"code":61,"score":22,"supported":14,"title":396,"abstract":397,"tags":398},[],[],[27],[400,401],"农业人工智能 植物病害识别 可解释性 智慧农业","农业人工智能 植物病害识别","农业人工智能植物病害识别可解释性智慧农业-5592","10.5281\u002Fzenodo.23545193",{"doi":403,"openalex_id":405,"authors":406,"venue":363,"cited_by_count":90,"oa_url":360,"card":413,"direction":169,"ingested_from":98},"W7223137323",[407,409,411],{"name":408,"orcid":9},"Akshay Ramgoolam",{"name":410,"orcid":9},"Priya Bundhoo Appadoo",{"name":412,"orcid":9},"Jean-Noël LEspérance",{"tldr":414,"method":415,"finding":416,"direction":169,"opportunity":417},"提出可解释性容量上界，评估低标签下概念秩崩溃审计的可靠性。","将分类器建模为概念秩算子，推导容量上界并做2000格功效模拟。","审计可靠性随概念秩谱与标签预算下降；5个百分点效应需52个聚类达80%功效。","可将容量上界用于小农户杀菌剂选择性决策的可解释模型审计与标签预算设计。","2026-10-22T11:21:10.425290Z",{"id":420,"title":421,"url":422,"summary":423,"summary_zh":424,"content":9,"source_name":363,"source_url":422,"published_at":364,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":365,"score_detail":425,"sources":428,"tags":430,"primary_topic":434,"secondary_topics":440,"classification":442,"search_phrases":465,"slug":468,"view_count":90,"doi":469,"paper":470,"created_at":482},5591,"Hierarchical Conformal Risk Control for Artificial Intelligence Agriculture Irrigation Scheduling Under Unmodeled Infiltration: A Computer Safety-Guarantee Evaluation","https:\u002F\u002Fdoi.org\u002F10.5281\u002Fzenodo.23543665","Irrigation scheduling algorithms that rely on soil-moisture forecasts are vulnerable to unmodeled infiltration, the process by which water bypasses the root zone through preferential pathways that resistance-based soil models cannot represent. The framework is evaluated through a replication-based design study that characterises the operating characteristics of the before-after-control-impact estimator used to audit scheduler performance, under a declared variance structure with between-cluster standard deviation 6.0, residual standard deviation 4.5 and intraclass correlation 0.64. Power was estimated across 2000 design cells spanning cluster counts of 8, 16, 24 and 32 and effect sizes of 3, 5 and 8 percentage points. The analysis indicates that 54 clusters are required to detect a 5-percentage-point effect at 80% power, and that the minimum detectable effect declines steeply as cluster count rises from 8 to 32. The contribution is a formal safety guarantee for irrigation decision support that remains valid when the infiltration model is wrong, together with an audit design that makes the guarantee empirically checkable.","依赖土壤湿度预报的灌溉调度算法容易受到未建模入渗的影响，即水分通过优先路径绕过根区的过程，而基于阻力的土壤模型无法表征这一过程。该框架通过一项基于重复的设计研究进行评估，该研究刻画了用于审计调度器性能的前后对照影响估计量的运行特性，其所依据的声明方差结构为：组间标准差6.0、残差标准差4.5、组内相关系数0.64。功效估计覆盖2000个设计单元，涵盖聚类数8、16、24和32以及效应量3、5和8个百分点。分析表明，在80%功效下检测5个百分点效应需要54个聚类，且最小可检测效应随聚类数从8增至32而急剧下降。本文的贡献在于为灌溉决策支持提供了一种正式的安全保证，该保证在入渗模型错误时仍然有效，同时提出了一种使该保证可经实证检验的审计设计。",{"impact":114,"substance":426,"depth":17,"authority":112,"freshness":184,"relevant":22,"comment":427},22,"该研究提出在入渗模型失配条件下仍有效的灌溉决策安全保证框架，并给出可审计的试验设计，方法新颖、结论可靠，但属细分方法学研究，产业影响有限，适合作为专业精选而非头条。",[429],{"name":363,"url":422},[27,28,431,432,433],"灌溉决策","安全保证","试验设计",{"code":435,"slug":436,"name":437,"path":438,"parent_code":439},"H3","irrigation","水资源与灌溉","\u002Ftopic\u002Fenvironment\u002Firrigation","H",[441],{"code":33,"slug":34,"name":35,"path":36,"parent_code":37},{"status":40,"locked":14,"version":41,"method":42,"evidence":443},{"candidates":444,"agricultural_evidence":47},[445,450,454,461],{"code":435,"score":113,"supported":47,"title":446,"abstract":447,"tags":449},[436],[448,436],"灌溉",[431],{"code":33,"score":112,"supported":47,"title":451,"abstract":452,"tags":453},[130],[245],[28],{"code":455,"score":21,"supported":47,"title":456,"abstract":457,"tags":460},"H2",[],[458,459],"土壤","soil",[],{"code":61,"score":22,"supported":14,"title":462,"abstract":463,"tags":464},[],[],[27],[466,467],"灌溉调度 安全保证","农业人工智能 安全保证 智慧农业 灌溉决策","灌溉调度安全保证-5591","10.5281\u002Fzenodo.23543665",{"doi":469,"openalex_id":471,"authors":472,"venue":363,"cited_by_count":90,"oa_url":422,"card":477,"direction":169,"ingested_from":98},"W7223131205",[473,475],{"name":474,"orcid":9},"Nassima Benkhelifa",{"name":476,"orcid":9},"Yacine Aït Ahmed",{"tldr":478,"method":479,"finding":480,"direction":169,"opportunity":481},"提出分层保形风险控制，为未建模入渗下的灌溉调度提供安全保证。","分层保形风险控制与基于重复设计的BACI审计，2000个设计单元功效分析。","检测5个百分点效应需54个集群，最小可检测效应随集群数增加陡降。","可将保形风险控制扩展到其他未建模扰动下的农业决策安全保证与审计设计。","2026-10-22T11:21:10.283561Z"]