[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-2997":3,"related-2997":45},{"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,"search_phrases":30,"slug":33,"view_count":34,"doi":8,"paper":35,"created_at":44},2997,"基于无人机多光谱图像和VGG21模型的小麦渍害调控效果识别方法","https:\u002F\u002Fwww.toutiao.com\u002Farticle\u002F7685741381196825088","江苏省农业科学院农业信息研究所梁万杰等联合中国农科院农业环境与可持续发展研究所、湖北粮作所、扬州大学等团队，针对小麦渍害防控提出基于无人机多光谱图像和VGG21模型的快速无损识别方法。在小麦拔节-抽穗和抽穗-灌浆两个阶段开展对照、渍水胁迫、硅肥调控和氨基酸调控4个类别数据集，大疆精灵4多光谱无人机采集小麦冠层多光谱图像，测产评估调控效果。",null,"智慧农业(中英文)2026,8(4):60-69","2026-09-15T12:42:00Z","论文",10,false,78,{"impact":16,"substance":17,"depth":16,"authority":18,"freshness":19,"relevant":20,"comment":21},18,22,14,6,1,"多机构协作提出无人机多光谱结合VGG21的小麦渍害无损识别方法，方法新颖、数据扎实，对智慧农业植保监测有参考价值。",[23],{"name":9,"url":6},[25,26,27,28,29],"智慧农业","农业人工智能","无人机遥感","小麦渍害","多光谱成像",[31,32],"江苏省农科院 小麦渍害 无人机多光谱","VGG21 小麦 渍害识别","江苏省农科院小麦渍害无人机多光谱-2997",0,{"doi":8,"openalex_id":8,"authors":36,"venue":8,"cited_by_count":34,"oa_url":8,"card":37,"direction":41,"ingested_from":43},[],{"tldr":38,"method":39,"finding":40,"direction":41,"opportunity":42},"用无人机多光谱图像和VGG21模型识别小麦渍害调控效果。","大疆精灵4多光谱无人机采集冠层图像，构建VGG21分类模型。","该方法可快速无损识别渍害及硅肥、氨基酸调控效果。","农业遥感与作物表型","可探索多光谱与深度学习结合评估其他逆境调控措施，并迁移至多作物场景。","agent","2026-09-20T00:03:07.858899Z",{"total":19,"page":20,"page_size":19,"items":46},[47,79,106,130,164,194],{"id":48,"title":49,"url":50,"summary":51,"summary_zh":8,"content":8,"source_name":52,"source_url":8,"published_at":53,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":54,"score_detail":55,"sources":62,"tags":64,"search_phrases":67,"slug":70,"view_count":34,"doi":8,"paper":71,"created_at":78},2854,"面向冬小麦水分含量的无人机遥感自动机器学习预测——MDPI Remote Sensing","https:\u002F\u002Fwww.mdpi.com\u002F2072-4292\u002F18\u002F18\u002F3161","本研究探索了无人机遥感快速准确评估冬小麦水分含量的潜力。在开花期和灌浆期使用配备多光谱、RGB和热红外相机的无人机获取高分辨率冠层遥感图像。集成地面真值采样数据与无人机遥感数据,使用自动机器学习(AutoML)框架建立回归模型预测冬小麦水分含量(MC)。结果表明,MC预测在灌浆期表现最佳,TIR传感器精度最高(R²=0.812,MAE=0.0204,RMSE=0.0274)。多传感器融合相比单传感器方法进一步提升预测性能,MC预测的R²达0.876、MAE 0.0191、RMSE 0.0259。来自中国农业科学院农田灌溉研究所。","MDPI Remote Sensing","2026-09-15T00:00:00Z",74,{"impact":56,"substance":57,"depth":58,"authority":59,"freshness":60,"relevant":20,"comment":61},15,21,17,13,8,"中国农科院团队用AutoML融合多光谱、RGB与热红外无人机数据预测冬小麦水分含量，多传感器融合R²达0.876，方法新颖、结论可靠，对精准灌溉有实用价值，值得进入每日精选。",[63],{"name":52,"url":50},[25,26,27,65,66],"冬小麦","多传感器融合",[68,69],"农业人工智能 多传感器融合 无人机遥感 智慧农业","农业人工智能 多传感器融合","农业人工智能多传感器融合无人机遥感智慧农业-2854",{"doi":8,"openalex_id":8,"authors":72,"venue":8,"cited_by_count":34,"oa_url":8,"card":73,"direction":41,"ingested_from":43},[],{"tldr":74,"method":75,"finding":76,"direction":41,"opportunity":77},"用无人机多传感器遥感结合AutoML预测冬小麦水分含量。","无人机多光谱、RGB、热红外图像+地面真值，AutoML回归建模。","灌浆期热红外精度最高，多传感器融合将R²提升至0.876。","可探索不同生育期与品种的泛化性，及将水分预测接入灌溉决策系统。","2026-09-18T00:03:30.758419Z",{"id":80,"title":81,"url":82,"summary":83,"summary_zh":8,"content":8,"source_name":84,"source_url":8,"published_at":85,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":86,"score_detail":87,"sources":89,"tags":91,"search_phrases":94,"slug":97,"view_count":34,"doi":8,"paper":98,"created_at":105},2611,"无人机遥感在水稻高通量表型分析中的研究进展 系统综述","https:\u002F\u002Fwww.ebiotrade.com\u002Fnewsf\u002F2026-9\u002F20260911171723051.htm","发表于Smart Agricultural Technology。依据PRISMA 2020规范系统检索文献最终纳入199项研究(2014–2026年)。研究发现：先进传感、特征集成和建模技术日益支持氮素和叶绿素估算及产量预测；轻量级模型和边缘计算系统在倒伏和病害监测任务中显示出实时部署的可行性；跨区域泛化受环境背景干扰以及地点品种偏倚制约；199项研究中有6项(3.0%)将UAV衍生性状与遗传关联分析联系起来。","Smart Agricultural Technology","2026-09-11T01:00:00Z",77,{"impact":16,"substance":17,"depth":16,"authority":59,"freshness":19,"relevant":20,"comment":88},"基于PRISMA规范纳入199项研究的系统综述，方法严谨、数据规模大，对水稻表型与智慧育种有实质参考价值，但属细分领域学术进展，公共影响有限。",[90],{"name":84,"url":82},[25,26,92,27,93],"水稻","高通量表型",[95,96],"农业人工智能 无人机遥感 高通量表型 智慧农业","农业人工智能 无人机遥感","农业人工智能无人机遥感高通量表型智慧农业-2611",{"doi":8,"openalex_id":8,"authors":99,"venue":8,"cited_by_count":34,"oa_url":8,"card":100,"direction":41,"ingested_from":43},[],{"tldr":101,"method":102,"finding":103,"direction":41,"opportunity":104},"系统综述199项研究，梳理无人机遥感在水稻高通量表型分析中的应用进展与瓶颈。","依据PRISMA 2020系统检索2014–2026年199项研究并归纳分析。","传感与建模支撑氮素、产量预测，跨区域泛化受环境与品种偏倚制约，基因关联研究仅占3%。","UAV表型与遗传关联分析严重不足，可探索跨区域泛化建模及表型-基因型融合方向。","2026-09-16T00:03:52.470505Z",{"id":107,"title":108,"url":109,"summary":110,"summary_zh":8,"content":111,"source_name":112,"source_url":8,"published_at":113,"category":114,"cover_url":8,"hotness":12,"is_selected":115,"score":116,"score_detail":117,"sources":121,"tags":123,"search_phrases":126,"slug":128,"view_count":20,"doi":8,"paper":8,"created_at":129},1499,"中国农大马韫韬教授团队 AI 驱动棉花高通量表型解析三项系列论文：覆盖花器官检测、单株表型、吐絮动态评估","https:\u002F\u002Fnews.cau.edu.cn\u002Fkxyj\u002F287ffb5af8464b76bd61d1db9d408af3.htm","中国农业大学土地科学与技术学院马韫韬教授团队围绕棉花关键生育时期的表型智能识别与量化分析，在 Computers and Electronics in Agriculture、Plant Phenomics 和 Precision Agriculture 等国际权威期刊上连续发表三篇研究论文，题目分别为 MS-R outperforms RGB: Lightweight CYOLO-BiCNet enables efficient detection and storage optimization of cotton flowers，Scalable phenotyping and yield estimation via stability index and single-plant variability using a vision-based large model framework 和 Assessing Cotton Boll-opening Concentration for Harvest Decision-making via Foundation Model-enhanced Cross-scale Phenotyping。","**中国农大新闻网讯**近日，中国农业大学土地科学与技术学院马韫韬教授团队在AI驱动的棉花高通量表型解析领域取得系列重要进展。团队围绕棉花关键生育时期的表型智能识别与量化分析，在 _Computers and Electronics in Agriculture_、_Plant Phenomics_ 和 _Precision Agriculture_ 等国际权威期刊上连续发表三篇研究论文，题目分别为[_MS-R outperforms RGB: Lightweight CYOLO-BiCNet enables efficient detection and storage optimization of cotton flowers_](https:\u002F\u002Flink.springer.com\u002Farticle\u002F10.1007\u002Fs11119-026-10396-9)_，[Scalable phenotyping and yield estimation via stability index and single-plant variability using a vision-based large model framework](https:\u002F\u002Fwww.sciencedirect.com\u002Fscience\u002Farticle\u002Fabs\u002Fpii\u002FS0168169925014061)_ 和[_Assessing Cotton Boll-opening Concentration for Harvest Decision-making via Foundation Model-enhanced Cross-scale Phenotyping_](https:\u002F\u002Fwww.sciencedirect.com\u002Fscience\u002Farticle\u002Fpii\u002FS264365152600107X)。三篇论文构建了覆盖“花器官检测—单株表型解析—吐絮动态评估”的棉花全生育期AI表型技术链条，为推动棉花智慧育种和机械化收获决策提供了重要的方法学支撑与技术方案。\n\n团队在作物表型组学与智能农业方向具有深厚的研究积累。近年来，又将研究视野从大田粮食作物拓展至经济作物领域，与中国农业科学院棉花研究所杜雄明研究员、何守朴研究员团队建立了紧密的合作关系。合作团队围绕棉花种质资源评价、株型调控与机械化适配等核心问题，联合开展了大量田间表型试验与数据采集工作，为AI表型技术的落地提供了丰富的多源、多尺度、多时段数据基础和育种应用场景。在此基础上，马韫韬教授团队充分发挥在深度学习建模、无人机遥感与高通量表型解析方面的技术优势，逐步形成了面向棉花全生育期的智能化表型研究体系。\n\n此次发表的三篇系列论文，分别从棉花不同生育阶段的关键表型需求出发，构建了一条由花器官快速检测、单株尺度多维表型解析到吐絮动态评估与收获决策的技术闭环。论文一针对棉花花期花朵检测的效率与存储瓶颈，提出了轻量化检测网络CYOLO-BiCNet，证实独立获取的多光谱红波段（MS-R）在花朵检测精度与存储效率上均显著优于传统RGB图像，为大田尺度花器官高通量监测提供了低成本、高效率的数据-模型组合方案。\n\n![Image 1: 图片1.png](https:\u002F\u002Fnews.cau.edu.cn\u002Fimages\u002F2026-09\u002F8ed95462d59c45a8ad68d69bb913d44e.png)\n\n图1 基于CYOLO-BiCNet模型对不同数据类型、多个时期棉花花的推断结果分析\n\n图1展示了CYOLO-BiCNet模型在MS-R、RGB及RGB-R三种数据上的可视化推理结果。整体来看，MS-R数据的漏检花朵数量最少，检测效果最优；RGB与 RGB-R数据漏检相对较多。从不同采集日期来看，7月30日三种数据的漏检数分别为11、16、19朵；8月8日降至4、6、7朵；8月15日为3、7、5朵。可视化结果直观表明，MS-R数据在各时段均表现出最佳的花朵检测性能，漏检情况显著少于RGB和RGB-R数据。\n\n![Image 2: 图片2.png](https:\u002F\u002Fnews.cau.edu.cn\u002Fimages\u002F2026-09\u002F31e657d8287840d198aff233328adebe.png)\n\n图2 基于视觉大模型的棉花单株多维表型解析与高产种质筛选技术流程\n\n论文二聚焦单株尺度表型解析这一育种核心需求，开发了视觉大模型框架TopoRefineSAM，融合YOLOv12检测与SAM2分割能力，实现了复杂田间条件下棉花单株的精准实例分割与多性状反演；同时创新性地提出稳定性指标组（SIG），将单株间的差异转化为小区尺度的稳定性特征，显著提升了产量估算和品种筛选的准确性与可解释性。\n\n![Image 3: 图片3.png](https:\u002F\u002Fnews.cau.edu.cn\u002Fimages\u002F2026-09\u002F76d4db4425564fd28a022cfd32166839.png)\n\n图3 有无稳定性指数组时，产量反演精度和特征重要性之间的差异\n\n图3展示了稳定性指数组（SIG）对小区产量反演精度的提升效果。引入SIG后，全生育期决定系数R²由0.344–0.629提升至0.512–0.701，RMSE与MAE同步下降。生育前期精度偏低，7月3日与10日R²分别由0.473、0.440升至0.512、0.520；随生育进程推进，模型表现持续改善，8月23日R²达全季最高0.701（无稳定性指数组时R²为0.629）。特征重要性分析显示，RHa_mean始终为最关键预测因子，而Ta_stability、Tl_stability、RHa_stability与 Tc_stability均进入前十，稳定性特征整体贡献超过61%。结果表明，SIG有效整合了株间变异信息，显著增强了模型对产量差异的解释能力与全生育期预测精度。\n\n![Image 4: 图片4.png](https:\u002F\u002Fnews.cau.edu.cn\u002Fimages\u002F2026-09\u002F1395532013ea4c71a7612cd15fbd7773.png)\n\n图4 棉花吐絮进程动态监测与基于CTSI指数的机械化收获决策框架\n\n论文三则瞄准棉花机械化收获中吐絮集中度评价这一产业痛点，开发了基于视觉基础模型的DINO-BollGX检测框架，结合两年期383个品种的多时相无人机影像，重建了小区尺度吐絮动态时序曲线，提出了棉花吐絮时间稳定性指数（CTSI），并将其与时间风险函数耦合生成收获决策曲线，为品种适宜机收性评价和最优收获窗口确定提供了定量化工具。\n\n![Image 5: 图片5.png](https:\u002F\u002Fnews.cau.edu.cn\u002Fimages\u002F2026-09\u002F6711154c55a84f5c8dd2f980c6a7f9e3.png)\n\n图5 383个品种棉花吐絮时空分布特征及代表性品种动态吐絮曲线聚类分析\n\n图5展示了多时相无人机棉铃吐絮计数的群体与品种尺度动态。群体尺度上，吐絮在时间分布上呈后期集中态势：观测初期（8月中下旬）各小区吐絮量极低，中位数均不足峰值的5%，箱线图四分位距狭窄，表明吐絮进程缓慢且表达微弱；进入9月中下旬后中位数快速攀升，四分位距显著拓宽，标志着吐絮进入加速期；10月14日至20日为吐絮高峰窗口，中位数由131跃升至277铃\u002F小区，约半数季节最大吐絮量在此短期内完成；11月初中位数小幅回落至244，提示吐絮基本完成并伴有少量损耗或遮挡。品种尺度上，383个品种整体表现为前期缓慢、后期快速增长的动态模式，但吐絮起始时间与后期增速差异显著。基于标准化动态特征的K-means聚类进一步将品种划分为两类：紧凑高量型吐絮窗口短（中位6天）、峰值高（中位460铃\u002F小区）、速率快（35铃\u002F天）；延续型窗口长（约60天）、峰值低（144铃\u002F小区）、速率缓（2.4铃\u002F天），两组差异均达极显著水平（p\u003C0.001）。\n\n三篇论文沿棉花“花—株—铃”的生育主线层层递进，形成了从器官识别、个体解析到群体决策的完整技术体系。这一系列成果的核心意义在于，将人工智能与高通量表型技术深度融合，为棉花育种从传统的经验选择向数据驱动的精准筛选转型提供了系统性解决方案。当前，我国棉花种质资源丰富但表型评价手段仍相对滞后，大量优异种质因缺乏高效、标准化的表型鉴定而未能被充分发掘利用。未来，团队将聚焦更多元化的棉花种质资源群体，结合新疆、河南等不同生态区的目标环境特征，拓展表型解析框架在不同棉区和种植制度下的适用性与泛化能力；同时，进一步整合数字孪生、作物生长模型与基因组信息，构建贯通“表型—基因型—环境—管理”的棉花智能设计体系，为我国棉花种业振兴和智慧农业的高质量发展提供核心技术支撑。\n\n上述三篇论文的第一作者均为土地科学与技术学院博士研究生陈勉，通讯作者为马韫韬教授。合作者包括中国农业科学院棉花研究所杜雄明研究员、何守朴研究员、耿晓丽副研究员、胡道武副研究员；中国农业科学院生物技术研究所张锐研究员；中国农业大学农学院田晓莉教授等，研究得到了国家重点研发计划、国家自然科学基金等项目的资助支持。","中国农业大学新闻网","2026-09-01T00:00:00Z","报道",true,89,{"impact":118,"substance":118,"depth":119,"authority":59,"freshness":60,"relevant":20,"comment":120},24,20,"中国农大团队在棉花AI表型领域发表三篇系列论文，构建全生育期技术链条，方法新颖、数据详实，对智慧育种和机收决策有重要支撑。",[122],{"name":112,"url":109},[25,26,124,27,125],"棉花","作物表型",[127,96],"农业人工智能 无人机遥感 作物表型 智慧农业","农业人工智能无人机遥感作物表型智慧农业-1499","2026-09-03T00:06:43.550281Z",{"id":131,"title":132,"url":133,"summary":134,"summary_zh":8,"content":8,"source_name":135,"source_url":133,"published_at":136,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":86,"score_detail":137,"sources":140,"tags":142,"search_phrases":144,"slug":146,"view_count":34,"doi":147,"paper":148,"created_at":163},1263,"Cognitive UAV-driven agro-surveillance framework for predicting crop stress–induced yield loss using spatio-temporal learning and adaptive irrigation control","https:\u002F\u002Fdoi.org\u002F10.1038\u002Fs41598-026-68174-6","Precision agriculture is becoming more and more of a challenge that requires the use of intelligent systems that are able to predict stress and prevent yield loss before it is too late. Traditional methods of agricultural surveillance are predominantly reactive with irrigation demands being based on thresholds or individual yield forecasts models that do not represent the intricate spatio-temporal interactions that exist between crop physiology, soil status, and environmental stresses. Besides, the majority of the current practices do not have an autonomous decision-making approach to preventive intervention which leads to inefficient use of water and slows down the response to stress. This paper suggests a cognitive UAV-assisted agro-surveillance system to predict yield vulnerability caused by crop stress and optimize adaptive irrigation with the help of spatio-temporal deep and reinforcement learning. The framework combines UAV-obtained RGB and multispectral and thermal imagery with measurements of soil sensors and meteorological data obtained with the Crop Health and Environmental Stress Dataset. A new GeoSpatio-TRiNet model is used to acquire long-range spatial relationship, time stress development, and diffusion of stresses across agricultural regions. The model predicts the vulnerability trajectories of the stress instead of the direct yield regression, and this allows early detection of yield risk. Such predictions serve to generate a cognitive environmental state of a Soft ActorCritic (SAC) reinforcement learning agent that autonomously computes zone-based irrigation behaviors to reduce the recurrence of stress at the minimum water usage cost. As shown by the results of the experiment, the proposed framework has a stress forecasting accuracy of 96.3% and performs much better than the traditional machine learning, CNN-based, and transformer-based baselines. The system also decreases the predicted yield vulnerability by 46.6 and enhances water-use efficiency by 41.1 as compared to irrigation strategies based on rules. The results confirm the usefulness of spatio-temporal intelligence with predictive control in terms of effectiveness, and the proposed framework is a scalable and sustainable solution to precision agriculture of the next generation.","Scientific Reports","2026-08-29T00:00:00Z",{"impact":119,"substance":17,"depth":16,"authority":18,"freshness":138,"relevant":20,"comment":139},3,"提出认知无人机农业监测框架，结合时空学习与强化学习预测作物胁迫并优化灌溉，精度高且节水显著，具前沿性。",[141],{"name":135,"url":133},[25,26,27,143],"精准灌溉",[145,96],"农业人工智能 无人机遥感 智慧农业 精准灌溉","农业人工智能无人机遥感智慧农业精准灌溉-1263","10.1038\u002Fs41598-026-68174-6",{"doi":147,"openalex_id":149,"authors":150,"venue":135,"cited_by_count":34,"oa_url":155,"card":156,"direction":160,"ingested_from":162},"W7204562418",[151,153],{"name":152,"orcid":8},"S. Selvakumar",{"name":154,"orcid":8},"D. Venugopal","https:\u002F\u002Fwww.nature.com\u002Farticles\u002Fs41598-026-68174-6_reference.pdf",{"tldr":157,"method":158,"finding":159,"direction":160,"opportunity":161},"提出认知无人机农业监测框架，用时空学习和强化学习预测作物胁迫并优化灌溉。","结合无人机多光谱热成像、土壤传感器和气象数据，用GeoSpatio-TRiNet","胁迫预测准确率96.3%，产量脆弱性降低46.6%，水分利用效率提升41.1%。","智慧农业 \u002F 农业物联网","可探索将框架扩展到多种作物和更大尺度，或集成实时决策与自主无人机路径规划。","openalex","2026-09-01T04:03:11.723322Z",{"id":165,"title":166,"url":167,"summary":168,"summary_zh":8,"content":8,"source_name":169,"source_url":8,"published_at":170,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":171,"score_detail":172,"sources":175,"tags":177,"search_phrases":181,"slug":184,"view_count":34,"doi":8,"paper":185,"created_at":193},3002,"改进生物神经网络的农业播种机全覆盖路径规划","https:\u002F\u002Fwww.mdpi.com\u002F2077-0472\u002F16\u002F18\u002F1968","江苏大学魏军等提出一种考虑播种与非播种运动模式切换机制的改进生物神经网络（BNN）方法，基于周围环境条件将下一节点状态分类为播种、封闭或转移节点。在BNN景观引导下机器沿平行直线路径继续播种操作；检测到封闭节点时切换至非播种模式并使用深度优先搜索算法搜索潜在封闭区域；检测到转移节点时同样切换非播种模式搜索合理的新目标节点。仿真表明该方法实现播种操作的完全覆盖同时避免重复遍历已播种区域。","MDPI Agriculture 16(18):1968","2026-09-14T00:00:00Z",69,{"impact":173,"substance":57,"depth":58,"authority":59,"freshness":19,"relevant":20,"comment":174},12,"提出改进生物神经网络的全覆盖路径规划方法，方法新颖、结论可靠，但属细分领域学术进展，公共影响有限。",[176],{"name":169,"url":167},[25,26,178,179,180],"智能农机","路径规划","播种机",[182,183],"江苏大学 播种机 全覆盖路径规划","生物神经网络 播种机 路径规划","江苏大学播种机全覆盖路径规划-3002",{"doi":8,"openalex_id":8,"authors":186,"venue":8,"cited_by_count":34,"oa_url":8,"card":187,"direction":191,"ingested_from":43},[],{"tldr":188,"method":189,"finding":190,"direction":191,"opportunity":192},"提出改进生物神经网络，实现农业播种机全覆盖路径规划并避免重复播种。","改进BNN结合节点分类与深度优先搜索，仿真验证。","方法实现播种完全覆盖，同时避免重复遍历已播种区域。","农业人工智能与决策模型","可结合真实农田地形与多机协同，验证动态环境下的路径规划鲁棒性。","2026-09-20T00:03:08.288198Z",{"id":195,"title":196,"url":197,"summary":198,"summary_zh":8,"content":8,"source_name":199,"source_url":8,"published_at":53,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":14,"score_detail":200,"sources":203,"tags":205,"search_phrases":208,"slug":211,"view_count":34,"doi":8,"paper":212,"created_at":219},3000,"Plant-GeoAT：几何感知3D植物点云器官身份解析用于器官级表型分析","https:\u002F\u002Fwww.mdpi.com\u002F2073-4395\u002F16\u002F18\u002F1809","四川农业大学吴俊杰等提出Plant-GeoAT几何感知解析网络，在RGB融合前编码局部相对-XYZ邻域并将空间邻域与特征空间关系耦合用于密集点预测。在自建油菜数据集、基于图像的大豆数据集、激光扫描Pheno4D玉米和番茄数据集上评估，5个随机种子下mIoU分别达92.26±0.28%、82.50±0.24%、99.74±0.05%、94.75±0.15%，玉米Stem IoU达99.57±0.09%。","MDPI Agronomy 16(18):1809",{"impact":16,"substance":201,"depth":16,"authority":59,"freshness":19,"relevant":20,"comment":202},23,"方法新颖、多作物多数据集验证且精度数据扎实，属器官级表型分析细分领域的重要技术进展，值得进入每日精选。",[204],{"name":199,"url":197},[25,26,206,93,207],"油菜","三维点云",[209,210],"四川农业大学 植物点云 器官识别","Plant-GeoAT 表型分析","四川农业大学植物点云器官识别-3000",{"doi":8,"openalex_id":8,"authors":213,"venue":8,"cited_by_count":34,"oa_url":8,"card":214,"direction":41,"ingested_from":43},[],{"tldr":215,"method":216,"finding":217,"direction":41,"opportunity":218},"提出几何感知网络Plant-GeoAT，实现3D植物点云器官身份解析与器官级表型分析。","编码局部相对XYZ邻域并耦合空间与特征空间关系，在油菜、大豆、玉米、番茄点云数据","五个数据集mIoU最高达99.74%，玉米茎IoU达99.57%，验证了几何感知对器官分割的有效性。","可探索跨物种、跨传感器的轻量化几何感知模型，并推动器官级表型与基因型关联分析。","2026-09-20T00:03:08.094512Z"]