[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-5613":3,"related-5613":94},{"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":23,"tags":25,"primary_topic":31,"secondary_topics":37,"classification":43,"search_phrases":68,"slug":71,"view_count":72,"doi":73,"paper":74,"created_at":93},5613,"ExplainableCropAI: Interpretable Deep Learning for Agricultural Disease Diagnosis and Decision Support","https:\u002F\u002Fdoi.org\u002F10.1007\u002Fs44196-026-01654-9","Deep learning models classify crop diseases accurately, but their black-box nature limits practical use: farmers and extension officers need to understand a diagnosis before acting on it. We present ExplainableCropAI, a decision-support system for paddy disease diagnosis that attaches three complementary explanation channels to a ConvNeXt V2-Tiny classifier: Grad-CAM++ visual attribution, Monte Carlo Dropout uncertainty with temperature-scaled calibration, and natural-language explanations generated by Phi-3-mini under a grounding constraint that restricts the model to typed fields drawn from the prediction, the attribution map and a curated agronomic knowledge base. None of the components is new; the contribution lies in their integration into a single pipeline and in measuring the properties that integration is meant to deliver. On a leakage-controlled, held-out test split of 1,540 images from the Paddy Doctor dataset (10,407 images, nine diseases and healthy), the classifier reaches 97.5% accuracy (95% CI 96.7–98.3) and macro-F1 of 96.9%. Six attribution methods are compared under identical settings with model-based faithfulness metrics; the calibrated confidence separates correct from incorrect predictions with an AUROC of 0.93, and referring the 10% least-confident cases to an expert removes 74% of all errors. Under eight image corruptions at five severities the model retains 89.7% mean accuracy against 72–78% for six baselines trained identically, and its referral rate rises with the degree of degradation. Zero-shot evaluation on two independently collected rice-disease datasets from Bangladesh and India shows that accuracy roughly halves under cross-dataset shift and that confidence no longer identifies errors, while fifty local images per class recover accuracy above 90%. The code, split files, trained weights and evaluation scripts are released.","深度学习模型能够准确分类作物病害，但其黑箱性质限制了实际应用：农民和推广人员需要在理解诊断结果后才能采取行动。我们提出ExplainableCropAI，一个用于水稻病害诊断的决策支持系统，它在ConvNeXt V2-Tiny分类器上附加了三条互补的解释通道：Grad-CAM++视觉归因、经温度缩放校准的蒙特卡洛Dropout不确定性，以及由Phi-3-mini在接地约束下生成的自然语言解释——该约束将模型限制为从预测结果、归因图和精选农学知识库中提取的类型化字段。各组件均非新创；贡献在于将其集成到单一流水线中，并测量该集成旨在提供的特性。在来自Paddy Doctor数据集（10,407张图像，九种病害及健康）的1,540张图像上，经泄漏控制的留出测试划分显示，分类器达到97.5%的准确率（95% CI 96.7–98.3）和96.9%的宏F1值。在相同设置下，使用基于模型的忠实性指标比较了六种归因方法；校准后的置信度以0.93的AUROC区分正确与错误预测，将置信度最低的10%案例转交专家可消除全部错误的74%。在五种严重程度的八种图像损坏下，模型保持89.7%的平均准确率，而相同训练的六个基线为72–78%，且其转交率随退化程度上升。在来自孟加拉国和印度的两个独立采集的水稻病害数据集上进行零样本评估显示，跨数据集偏移下准确率大约减半，置信度不再能识别错误，而每类五十张本地图像可将准确率恢复至90%以上。代码、划分文件、训练权重和评估脚本均已发布。",null,"International Journal of Computational Intelligence Systems","2026-10-22T00:00:00Z","论文",10,false,84,{"impact":17,"substance":18,"depth":19,"authority":20,"freshness":13,"relevant":21,"comment":22},18,23,19,14,1,"可解释AI与作物病害诊断的集成系统，方法组合有工程价值，数据规模与评测较扎实，但组件本身非新，属细分领域重要进展而非重大突破。",[24],{"name":10,"url":6},[26,27,28,29,30],"智慧农业","农业人工智能","水稻","可解释AI","作物病害诊断",{"code":32,"slug":33,"name":34,"path":35,"parent_code":36},"I2","agricultural-ai","农业AI与大模型","\u002Ftopic\u002Fdigital-agriculture\u002Fagricultural-ai","I",[38],{"code":39,"slug":40,"name":28,"path":41,"parent_code":42},"A1","rice","\u002Ftopic\u002Ffield-crops\u002Frice","A",{"status":44,"locked":14,"version":45,"method":46,"evidence":47},"classified","2026-10-22.2","rules",{"candidates":48,"agricultural_evidence":51},[49,58,63],{"code":32,"score":50,"supported":51,"title":52,"abstract":55,"tags":57},24,true,[53,54],"deep learning","decision support",[56,53],"深度学习",[27,29],{"code":39,"score":13,"supported":51,"title":59,"abstract":60,"tags":62},[],[28,40,61],"paddy",[28],{"code":64,"score":21,"supported":14,"title":65,"abstract":66,"tags":67},"I1",[],[],[26],[69,70],"ExplainableCropAI 水稻病害","Paddy Doctor 数据集","ExplainableCropAI水稻病害-5613",0,"10.1007\u002Fs44196-026-01654-9",{"doi":73,"openalex_id":75,"authors":76,"venue":10,"cited_by_count":72,"oa_url":6,"card":86,"direction":90,"ingested_from":92},"W7166554515",[77,79,80,83,85],{"name":78,"orcid":9},"Santosh Kumar",{"name":78,"orcid":9},{"name":81,"orcid":82},"Durgesh Nandan","https:\u002F\u002Forcid.org\u002F0000-0002-9762-3559",{"name":84,"orcid":9},"Sumit Kumar",{"name":84,"orcid":9},{"tldr":87,"method":88,"finding":89,"direction":90,"opportunity":91},"构建可解释水稻病害诊断系统，融合归因、不确定性与自然语言解释。","ConvNeXt V2-Tiny分类器，Grad-CAM++、MC Dropou","测试准确率97.5%，低置信转诊消除74%错误，跨数据集精度减半。","农业人工智能与决策模型","跨数据集置信失效，少量本地样本微调可恢复精度，值得研究田间自适应与主动学习。","openalex","2026-10-22T23:31:14.582261Z",{"total":95,"page":21,"page_size":95,"items":96},6,[97,173,232,314,373,459],{"id":98,"title":99,"url":100,"summary":101,"summary_zh":102,"content":9,"source_name":103,"source_url":100,"published_at":104,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":105,"score_detail":106,"sources":111,"tags":113,"primary_topic":116,"secondary_topics":117,"classification":119,"search_phrases":141,"slug":144,"view_count":21,"doi":145,"paper":146,"created_at":172},4993,"Effects of within-season temporal aggregation of weather variables on performance and interpretability in rice yield machine learning models","https:\u002F\u002Fdoi.org\u002F10.1038\u002Fs41598-026-75646-2","While crop responses are inherently stage-dependent, machine learning yield models often rely on weather variables aggregated over the entire growing season. This study evaluates how within-season temporal aggregation of weather variables affects both predictive performance and interpretability in machine learning-based rice yield modeling. Using CYGMA (Crop Yield Growth Model with Assumptions on Climate and Socioeconomics)-simulated potential yield as a synthetic response variable, we minimized confounding effects from observational noise and management variability, enabling a clearer evaluation of the role of within-season weather aggregation. Daily weather variables were averaged over the rice growing period after dividing each season into multiple equal-length segments, and several machine learning algorithms were trained and evaluated using temporally independent validation. Increasing within-season temporal resolution improved model performance up to a moderate level, after which gains diminished. Interpretation using SHAP (SHapley Additive exPlanations) and response diagnostics consistently indicated that yield sensitivity concentrates in specific windows of the season, with temperature-related effects emerging most strongly early in the season and humidity-related effects later. Finer temporal aggregation sharpened the identification of window-specific nonlinear and threshold-like responses that were obscured under coarse seasonal averages. Overall, moderate within-season aggregation provides a practical balance between predictive skill and mechanistic interpretability, supporting more reliable identification of phenology-relevant meteorological risks for climate-impact assessment and decision support.","尽管作物响应本质上具有阶段依赖性，但机器学习产量模型往往依赖于整个生长季聚合的天气变量。本研究评估了天气变量的季节内时间聚合如何影响基于机器学习的水稻产量建模的预测性能与可解释性。以CYGMA（Crop Yield Growth Model with Assumptions on Climate and Socioeconomics，气候与社会经济假设下的作物产量生长模型）模拟的潜在产量作为合成响应变量，我们最大限度地减少了观测噪声和管理变异性带来的混杂效应，从而能够更清晰地评估季节内天气聚合的作用。将每个生长季划分为多个等长时段后，对水稻生长期内的逐日天气变量进行平均，并采用时间独立验证对多种机器学习算法进行训练和评估。提高季节内时间分辨率可提升模型性能，直至达到中等水平，此后增益递减。利用SHAP（SHapley Additive exPlanations，沙普利加性解释）和响应诊断进行的解释一致表明，产量敏感性集中在季节的特定窗口，温度相关效应在季节早期最为显著，湿度相关效应则在后期出现。更精细的时间聚合增强了对窗口特异性非线性和阈值型响应的识别，而这些响应在粗粒度季节平均下被掩盖。总体而言，中等程度的季节内聚合在预测能力与机制可解释性之间提供了实用平衡，有助于更可靠地识别与物候相关的气象风险，以支持气候影响评估和决策支持。","Scientific Reports","2026-10-13T00:00:00Z",72,{"impact":107,"substance":108,"depth":17,"authority":20,"freshness":109,"relevant":21,"comment":110},12,20,8,"该研究通过CYGMA模拟数据系统评估了生长季内气象变量时间聚合对水稻产量机器学习模型性能与可解释性的影响，方法新颖、结论可靠，对农业气象风险识别与决策支持有参考价值，但属学术论文，公共影响范围有限。",[112],{"name":103,"url":100},[26,27,29,114,115],"农业气象","水稻产量预测",{"code":39,"slug":40,"name":28,"path":41,"parent_code":42},[118],{"code":32,"slug":33,"name":34,"path":35,"parent_code":36},{"status":44,"locked":14,"version":45,"method":46,"evidence":120},{"candidates":121,"agricultural_evidence":51},[122,128,132,137],{"code":32,"score":19,"supported":51,"title":123,"abstract":125,"tags":127},[124],"machine learning",[126,124,54],"机器学习",[27,29],{"code":39,"score":20,"supported":51,"title":129,"abstract":130,"tags":131},[40],[28,40],[],{"code":133,"score":21,"supported":14,"title":134,"abstract":135,"tags":136},"H7",[],[],[114],{"code":64,"score":21,"supported":14,"title":138,"abstract":139,"tags":140},[],[],[26],[142,143],"水稻产量 机器学习 气象","SHAP 水稻 产量预测","水稻产量机器学习气象-4993","10.1038\u002Fs41598-026-75646-2",{"doi":145,"openalex_id":147,"authors":148,"venue":103,"cited_by_count":72,"oa_url":165,"card":166,"direction":171,"ingested_from":92},"W7221657053",[149,151,153,155,157,159,162],{"name":150,"orcid":9},"Noh-Hyun Lee",{"name":152,"orcid":9},"Wonjae Jeong",{"name":154,"orcid":9},"Hyeon-Ji Yang",{"name":156,"orcid":9},"Hana Cho",{"name":158,"orcid":9},"Hyeon-Yeong Ju",{"name":160,"orcid":161},"Toshichika Iizumi","https:\u002F\u002Forcid.org\u002F0000-0002-0611-4637",{"name":163,"orcid":164},"Ki‐Tae Kim","https:\u002F\u002Forcid.org\u002F0000-0002-8249-5283","https:\u002F\u002Fwww.nature.com\u002Farticles\u002Fs41598-026-75646-2_reference.pdf",{"tldr":167,"method":168,"finding":169,"direction":90,"opportunity":170},"评估生长季内天气变量时间聚合粒度对水稻产量机器学习模型预测与可解释性的影响。","用CYGMA模拟潜在产量作响应变量，按等长时段聚合逐日天气，训练多种机器学习模型","适度提高季内时间分辨率可提升性能，产量敏感性集中在特定窗口，温度效应偏早、湿度效应偏晚。","可探索不同作物与气候区下最优聚合粒度，并将窗口特异性气象风险嵌入气候影响评估与决策支持系统。","智慧农业 \u002F 农业物联网","2026-10-15T23:30:10.483881Z",{"id":174,"title":175,"url":176,"summary":177,"summary_zh":178,"content":9,"source_name":179,"source_url":176,"published_at":180,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":181,"score_detail":182,"sources":187,"tags":189,"primary_topic":192,"secondary_topics":193,"classification":195,"search_phrases":211,"slug":214,"view_count":72,"doi":215,"paper":216,"created_at":231},4802,"A Privacy-Preserving Federated Learning Framework for Paddy Leaf Disease Detection Using an Enhanced Pretrained Novel EfficientNet-Lite4 Model","https:\u002F\u002Fdoi.org\u002F10.15832\u002Fankutbd.1837705","The correct and prompt detection of leaf diseases in paddy crops remains essential for protecting agricultural output and yield stability and reducing major losses that occur during severe disease outbreaks. The widespread use of conventional centralized machine learning techniques becomes impractical for real-world agricultural applications, as their effectiveness diminishes under three main conditions: data privacy limitations, delayed transmission of information, along with reduced performance accuracy from non-independent and identically distributed (non-IID) data properties across different agro-climatic regions. This research develops a decentralized diagnostic system which combines Federated Learning with an Enhanced Pretrained Novel EfficientNet-Lite4 model designed for paddy leaf disease classification. The Paddy Doctor dataset served as training material for the system which contained 16,225 RGB images that classified 12 disease types alongside healthy categories. The proposed model reached 98.7% classification accuracy and 98.5% precision and 98.6% F1-score while maintaining a final loss of 0.21 after 100 communication rounds. The system underwent field testing which demonstrated its ability to generate accurate diagnostic results in real time in various working conditions. The study demonstrates how privacy preserving AI-based paddy disease management works in practice and recommends local agricultural institutions to build decentralized learning systems. The suggested system will give farmers direct access to field level decision support and facilitate the creation of sustainable digital pest management systems required to ensure food supply.","正确且及时地检测水稻作物叶片病害，对于保护农业产出和产量稳定、减少病害严重暴发期间发生的重大损失仍然至关重要。传统集中式机器学习技术的广泛使用在现实农业应用中变得不切实际，因为在三种主要条件下其有效性会下降：数据隐私限制、信息传输延迟，以及不同农业气候区域间非独立同分布（non-IID）数据特性导致的性能准确率降低。本研究开发了一种去中心化诊断系统，将联邦学习与一种增强的预训练新型高效网络模型EfficientNet-Lite4相结合，用于水稻叶片病害分类。Paddy Doctor数据集作为该系统的训练材料，包含16,225张RGB图像，涵盖12种病害类型以及健康类别。所提出的模型在100轮通信后达到了98.7%的分类准确率、98.5%的精确率和98.6%的F1分数，同时最终损失保持在0.21。该系统经过实地测试，证明其能够在各种工作条件下实时生成准确的诊断结果。该研究展示了基于隐私保护人工智能的水稻病害管理如何在实践中运作，并建议地方农业机构构建去中心化学习系统。所建议的系统将为农民提供直接的田间决策支持，并促进建立确保粮食供应所需的可持续数字病虫害管理系统。","Tarım Bilimleri Dergisi","2026-10-12T00:00:00Z",80,{"impact":17,"substance":183,"depth":17,"authority":184,"freshness":185,"relevant":21,"comment":186},22,13,9,"将联邦学习与轻量级EfficientNet-Lite4结合用于水稻叶病识别，兼顾隐私保护与98.7%高精度，对智慧农业落地有参考价值。",[188],{"name":179,"url":176},[26,27,28,190,191],"病虫害识别","联邦学习",{"code":39,"slug":40,"name":28,"path":41,"parent_code":42},[194],{"code":32,"slug":33,"name":34,"path":35,"parent_code":36},{"status":44,"locked":14,"version":45,"method":46,"evidence":196},{"candidates":197,"agricultural_evidence":51},[198,203,207],{"code":39,"score":199,"supported":51,"title":200,"abstract":201,"tags":202},15,[61],[28,61],[28],{"code":32,"score":107,"supported":51,"title":204,"abstract":205,"tags":206},[],[126,124,54],[27,190,191],{"code":64,"score":21,"supported":14,"title":208,"abstract":209,"tags":210},[],[],[26],[212,213],"Paddy Doctor 数据集 水稻病害","EfficientNet-Lite4 水稻叶片病害","PaddyDoctor数据集水稻病害-4802","10.15832\u002Fankutbd.1837705",{"doi":215,"openalex_id":217,"authors":218,"venue":179,"cited_by_count":72,"oa_url":225,"card":226,"direction":171,"ingested_from":92},"W7221522859",[219,222],{"name":220,"orcid":221},"B Johnson","https:\u002F\u002Forcid.org\u002F0009-0009-8281-4051",{"name":223,"orcid":224},"Chandrakumar Thangavel","https:\u002F\u002Forcid.org\u002F0000-0002-1186-5988","https:\u002F\u002Fdergipark.org.tr\u002Fen\u002Fdownload\u002Farticle-file\u002F5486058",{"tldr":227,"method":228,"finding":229,"direction":90,"opportunity":230},"提出联邦学习结合改进EfficientNet-Lite4的水稻叶片病害检测框架，兼顾隐私与精度。","联邦学习+预训练EfficientNet-Lite4，用Paddy Doctor","12类病害分类准确率达98.7%，100轮通信后损失0.21，支持实时诊断。","可探索非IID数据下联邦聚合优化与轻量化模型在边缘设备的部署研究。","2026-10-13T23:30:08.809781Z",{"id":233,"title":234,"url":235,"summary":236,"summary_zh":9,"content":9,"source_name":237,"source_url":235,"published_at":238,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":239,"score_detail":240,"sources":242,"tags":244,"primary_topic":247,"secondary_topics":248,"classification":255,"search_phrases":275,"slug":278,"view_count":21,"doi":279,"paper":280,"created_at":313},4500,"Integrating UAV-Based Remote Sensing and Deep Learning for Early Detection of Bacterial Leaf Blight in Rice","https:\u002F\u002Fdoi.org\u002F10.21203\u002Frs.3.rs-11163942\u002Fv1","Integrating UAV-Based Remote Sensing and Deep Learning for Early Detection of Bacterial Leaf Blight in Rice。Research Square","Research Square","2026-10-08T00:00:00Z",74,{"impact":17,"substance":108,"depth":17,"authority":13,"freshness":109,"relevant":21,"comment":241},"将无人机遥感与深度学习结合用于水稻白叶枯病早期检测，方法有创新、结论具应用价值，但属预印本且为细分技术进展，适合进入主题聚合而非头条精选。",[243],{"name":237,"url":235},[26,27,28,245,246],"遥感","病害检测",{"code":39,"slug":40,"name":28,"path":41,"parent_code":42},[249,250],{"code":32,"slug":33,"name":34,"path":35,"parent_code":36},{"code":251,"slug":252,"name":253,"path":254,"parent_code":36},"I3","remote-sensing","遥感与农情监测","\u002Ftopic\u002Fdigital-agriculture\u002Fremote-sensing",{"status":44,"locked":14,"version":45,"method":46,"evidence":256},{"candidates":257,"agricultural_evidence":51},[258,262,266,271],{"code":39,"score":107,"supported":51,"title":259,"abstract":260,"tags":261},[40],[40],[28],{"code":32,"score":107,"supported":51,"title":263,"abstract":264,"tags":265},[53],[53],[27],{"code":251,"score":107,"supported":51,"title":267,"abstract":269,"tags":270},[268],"remote sensing",[268],[245],{"code":64,"score":21,"supported":14,"title":272,"abstract":273,"tags":274},[],[],[26],[276,277],"无人机遥感 水稻 白叶枯病","深度学习 水稻病害 早期检测","无人机遥感水稻白叶枯病-4500","10.21203\u002Frs.3.rs-11163942\u002Fv1",{"doi":279,"openalex_id":281,"authors":282,"venue":237,"cited_by_count":72,"oa_url":235,"card":9,"direction":312,"ingested_from":92},"W7221055917",[283,286,289,291,293,296,298,300,303,305,307,309],{"name":284,"orcid":285},"Carlito B. Balingbing","https:\u002F\u002Forcid.org\u002F0000-0002-3672-9253",{"name":287,"orcid":288},"Dmytro Chebotarov","https:\u002F\u002Forcid.org\u002F0000-0003-1351-9453",{"name":290,"orcid":9},"Ian  Carlo Perilla",{"name":292,"orcid":9},"Stephen Klassen",{"name":294,"orcid":295},"Nancy Castilla","https:\u002F\u002Forcid.org\u002F0000-0002-9143-4933",{"name":297,"orcid":9},"Raphael  Gabriel Carvajal",{"name":299,"orcid":9},"Kendrix Ortega",{"name":301,"orcid":302},"Joseph M. Sandro","https:\u002F\u002Forcid.org\u002F0000-0002-8311-2299",{"name":304,"orcid":9},"Nurmi Pangesti",{"name":306,"orcid":9},"Caesar Arloo R. Centeno",{"name":308,"orcid":9},"Arnel Rala",{"name":310,"orcid":311},"Li-Yu  Daisy Liu","https:\u002F\u002Forcid.org\u002F0000-0001-6997-8101","农业遥感与作物表型","2026-10-09T23:30:18.390549Z",{"id":315,"title":316,"url":317,"summary":318,"summary_zh":319,"content":9,"source_name":103,"source_url":317,"published_at":320,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":321,"score_detail":322,"sources":324,"tags":326,"primary_topic":329,"secondary_topics":330,"classification":332,"search_phrases":353,"slug":356,"view_count":357,"doi":358,"paper":359,"created_at":372},4431,"A multimodal deep learning for rice crop health monitoring using vegetation index feature selection","https:\u002F\u002Fdoi.org\u002F10.1038\u002Fs41598-026-74895-5","Crop health monitoring is essential for sustaining agricultural productivity and ensuring stable production, particularly in major crop-producing regions that face challenges due to climate variability, resource constraints, and spatial heterogeneity. Traditional methods struggle to effectively integrate spatial heterogeneity and environmental variability from multi-source remote sensing (RS) data. To address this problem, the proposed work explores the role of satellite imagery in monitoring rice crop health with the help of Vegetation Indices (VIs) under the impact of changing weather conditions. Multi-temporal Sentinel-1 and Sentinel-2 satellite data were acquired from a rice cultivation field in Gorakhpur, Uttar Pradesh, India, during three Kharif growing seasons (2023-2025). In this work, we propose a supervised MI-based feature selection framework employing Otsu’s adaptive thresholding for selection of VIs that strongly determine the health of rice crop. To ensure data integrity, cloud-contaminated observations were systematically excluded through a two-stage cleaning process. Further, we propose a multimodal deep learning (DL) framework that fuses a Swin Transformer (Swin-T) backbone for spatial feature extraction from vegetation index imagery with a Multi-Layer Perceptron (MLP) for structured numerical modeling of meteorological variables. The proposed framework achieved 94.05% accuracy on an independent held-out test set and a mean accuracy of \\(92.12 \\pm 5.28\\) % under 10-fold cross-validation, consistently outperforming competing hybrid deep learning baselines. These results highlight that integrating transformer-based spatial representations with environmental information provides an efficient approach for rice crop health classification. The proposed framework can be extended to multi-region crop monitoring by incorporating additional vegetation index modalities and diverse climatic conditions.","作物健康监测对于维持农业生产力、保障稳定生产至关重要，尤其是在面临气候变异性、资源约束和空间异质性挑战的主要作物产区。传统方法难以有效整合多源遥感（RS）数据中的空间异质性和环境变异性。为解决这一问题，本研究探索了卫星影像在天气条件变化影响下借助植被指数（VIs）监测水稻作物健康的作用。研究在印度北方邦戈勒克布尔的稻田中，于三个水稻生长季（2023—2025年）获取了多时相Sentinel-1和Sentinel-2卫星数据。在本研究中，我们提出了一种基于互信息（MI）的监督式特征选择框架，采用Otsu自适应阈值法筛选对水稻作物健康具有强判别能力的植被指数。为确保数据完整性，通过两阶段清洗流程系统性地剔除了受云污染的观测数据。此外，我们提出了一种多模态深度学习（DL）框架，该框架融合了Swin Transformer（Swin-T）主干网络用于从植被指数影像中提取空间特征，以及多层感知机（MLP）用于气象变量的结构化数值建模。所提框架在独立留出测试集上达到了94.05%的准确率，在10折交叉验证下平均准确率为\\(92.12 \\pm 5.28\\)%，持续优于竞争的混合深度学习基线模型。这些结果表明，将基于Transformer的空间表征与环境信息相结合，为水稻作物健康分类提供了一种高效方法。所提框架可通过纳入更多植被指数模态和多样化气候条件，扩展至多区域作物监测。","2026-10-07T00:00:00Z",81,{"impact":17,"substance":183,"depth":17,"authority":20,"freshness":185,"relevant":21,"comment":323},"方法新颖、数据跨三个生长季且精度可靠，属农业遥感与AI交叉的细分领域进展，值得进入每日精选。",[325],{"name":103,"url":317},[26,27,28,327,328],"遥感监测","作物长势",{"code":39,"slug":40,"name":28,"path":41,"parent_code":42},[331],{"code":32,"slug":33,"name":34,"path":35,"parent_code":36},{"status":44,"locked":14,"version":45,"method":46,"evidence":333},{"candidates":334,"agricultural_evidence":51},[335,340,344,349],{"code":39,"score":17,"supported":51,"title":336,"abstract":337,"tags":339},[40],[28,338,40],"稻田",[28],{"code":32,"score":199,"supported":51,"title":341,"abstract":342,"tags":343},[53],[56,53],[27],{"code":251,"score":345,"supported":51,"title":346,"abstract":347,"tags":348},7,[],[245,268],[327],{"code":64,"score":21,"supported":14,"title":350,"abstract":351,"tags":352},[],[],[26],[354,355],"Sentinel-2 水稻 遥感监测","Swin Transformer 作物健康","Sentinel-2水稻遥感监测-4431",2,"10.1038\u002Fs41598-026-74895-5",{"doi":358,"openalex_id":360,"authors":361,"venue":103,"cited_by_count":72,"oa_url":317,"card":367,"direction":312,"ingested_from":92},"W7220838898",[362,365],{"name":363,"orcid":364},"Shashank Gupta","https:\u002F\u002Forcid.org\u002F0000-0003-1291-7951",{"name":366,"orcid":9},"Shefali Arora Chouhan",{"tldr":368,"method":369,"finding":370,"direction":312,"opportunity":371},"融合Sentinel-1\u002F2植被指数与气象数据，用多模态深度学习监测水稻健康。","Sentinel-1\u002F2多时相数据、MI特征选择、Otsu阈值、Swin-T+M","独立测试准确率94.05%，10折交叉验证92.12%，优于混合基线。","可扩展至多区域多作物，探索更多植被指数模态与极端气候下的泛化能力。","2026-10-08T23:30:29.254071Z",{"id":374,"title":375,"url":376,"summary":377,"summary_zh":378,"content":9,"source_name":379,"source_url":376,"published_at":380,"category":12,"cover_url":9,"hotness":381,"is_selected":14,"score":105,"score_detail":382,"sources":386,"tags":390,"primary_topic":392,"secondary_topics":393,"classification":396,"search_phrases":443,"slug":446,"view_count":72,"doi":447,"paper":448,"created_at":458},4310,"Comprehensive remote sensing and machine learning framework for multi-regional rice yield forecasting under hydroclimatic variability in Sahelian Niger","https:\u002F\u002Fdoi.org\u002F10.5281\u002Fzenodo.23149579","It is crucial to forecast rice yields effectively in order to improve agricultural planning and resilience to climate change in the Sahel, where production is greatly affected by hydroclimatic variability. This research has developed an integrated framework combining remote sensing analysis and machine learning to assess the variability in rice production in Niamey, Dosso, Tillaberi and Diffa, in Niger, over the period 2010–2025. They combined Landsat-derived NDVI and EVI with parameters such as evapotranspiration, precipitation, temperature, humidity, solar radiation, soil moisture, cultivated area, year and season. We processed and harmonised the datasets using Google Earth Engine and Python. We then evaluated four regression algorithms separately, Random Forest (RF), k-nearest neighbours (KNN), support vector regression (SVR) and XGBoost (XGB), for each region. The correlation analysis revealed modest positive links between rice yield and the NDVI (r = 0.27) and the EVI (r = 0.14), while a strong correlation was observed between the NDVI and the EVI (r = 0.80). The vegetation indices showed marked seasonal and interannual variability, generally exhibiting higher values during the wet season. Model performance varied considerably from one region to another","在萨赫勒地区，有效预测水稻产量对于改善农业规划和增强气候变化适应能力至关重要，因为该地区的生产在很大程度上受到水文气候变率的制约。本研究开发了一个集成框架，将遥感分析与机器学习相结合，以评估尼日尔尼亚美、多索、蒂拉贝里和迪法在2010—2025年期间水稻产量的变异性。研究将基于Landsat的归一化植被指数（NDVI）和增强型植被指数（EVI）与蒸散发、降水、温度、湿度、太阳辐射、土壤水分、耕地面积、年份和季节等参数相结合。利用Google Earth Engine和Python对数据集进行了处理和协调。随后，针对每个地区分别评估了四种回归算法：随机森林（RF）、k近邻（KNN）、支持向量回归（SVR）和XGBoost（XGB）。相关分析显示，水稻产量与NDVI（r = 0.27）和EVI（r = 0.14）之间存在弱正相关，而NDVI与EVI之间则观察到强相关（r = 0.80）。植被指数表现出明显的季节性和年际变异性，通常在湿季呈现较高值。模型性能在不同地区之间差异显著。","Zenodo (CERN European Organization for Nuclear Research)","2026-10-05T00:00:00Z",25,{"impact":107,"substance":383,"depth":384,"authority":184,"freshness":185,"relevant":21,"comment":385},21,17,"基于Landsat植被指数与四种机器学习模型的多区域水稻产量预测研究，方法扎实、数据跨度长，但属区域性学术成果，对国内农业信息化有方法借鉴价值。",[387,388],{"name":379,"url":376},{"name":379,"url":389},"https:\u002F\u002Fdoi.org\u002F10.5281\u002Fzenodo.23149580",[26,27,28,391,245],"产量预测",{"code":39,"slug":40,"name":28,"path":41,"parent_code":42},[394,395],{"code":251,"slug":252,"name":253,"path":254,"parent_code":36},{"code":32,"slug":33,"name":34,"path":35,"parent_code":36},{"status":44,"locked":14,"version":45,"method":46,"evidence":397},{"candidates":398,"agricultural_evidence":51},[399,404,408,412,419,426,432,439],{"code":251,"score":400,"supported":51,"title":401,"abstract":402,"tags":403},16,[268],[245,268],[391,245],{"code":39,"score":199,"supported":51,"title":405,"abstract":406,"tags":407},[40],[28,40],[28],{"code":32,"score":199,"supported":51,"title":409,"abstract":410,"tags":411},[124],[126,124],[27],{"code":413,"score":95,"supported":51,"title":414,"abstract":415,"tags":418},"H2",[],[416,417],"土壤","soil",[],{"code":420,"score":95,"supported":51,"title":421,"abstract":422,"tags":425},"H3",[],[423,424],"蒸散发","evapotranspiration",[],{"code":133,"score":95,"supported":51,"title":427,"abstract":428,"tags":431},[],[429,430],"气候变化","climate change",[],{"code":433,"score":434,"supported":51,"title":435,"abstract":436,"tags":438},"J1",3,[],[437],"农业规划",[],{"code":64,"score":21,"supported":14,"title":440,"abstract":441,"tags":442},[],[],[26],[444,445],"尼日尔 水稻 遥感 产量预测","Landsat NDVI 水稻 机器学习","尼日尔水稻遥感产量预测-4310","10.5281\u002Fzenodo.23149579",{"doi":447,"openalex_id":449,"authors":450,"venue":379,"cited_by_count":72,"oa_url":376,"card":453,"direction":312,"ingested_from":92},"W7220391503",[451],{"name":452,"orcid":9},"Battula Vijaya Saradhi Halarou Saidou Souley",{"tldr":454,"method":455,"finding":456,"direction":312,"opportunity":457},"结合遥感植被指数与机器学习，预测尼日尔萨赫勒四地区2010–2025年水稻产量。","Landsat NDVI\u002FEVI 加气象土壤变量，GEE与Python处理，比较","NDVI\u002FEVI与产量仅弱相关，模型精度区域差异大，植被指数呈明显季节与年际波动。","可引入时序深度学习与迁移学习，解决跨区域泛化差及植被指数解释力弱的问题。","2026-10-06T23:30:20.077977Z",{"id":460,"title":461,"url":462,"summary":463,"summary_zh":464,"content":9,"source_name":465,"source_url":462,"published_at":466,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":181,"score_detail":467,"sources":469,"tags":471,"primary_topic":473,"secondary_topics":474,"classification":477,"search_phrases":499,"slug":502,"view_count":357,"doi":503,"paper":504,"created_at":520},3164,"Spatiotemporal Deep Learning for Rice Plant Height Estimation from Multi-Temporal UAV RGB Imagery","https:\u002F\u002Fdoi.org\u002F10.3390\u002Fagriculture16182034","Accurate plant height estimation is important for monitoring crop growth and supporting precision agricultural management. Manual measurements are labor-intensive, while LiDAR-based methods are expensive and require complex processing. UAV photogrammetry provides a lower-cost alternative but remains challenging in flooded rice paddies because of canopy deformation and difficulties in terrain extraction. This study proposes Rice-STNet, a spatiotemporal deep learning framework for end-to-end rice plant height estimation using multi-temporal UAV RGB imagery. Rice-STNet integrates a convolutional neural network for spatial feature extraction, Time2Vec for temporal encoding, and a gated recurrent unit network for modeling temporal dependencies across observation dates. The framework was evaluated using field data collected from rice paddies over two growing seasons. Rice-STNet achieved an R2 of 0.97, a root mean squared error of 1.97 cm, and a mean absolute error of 1.14 cm. It outperformed random forest, support vector regression, a CNN-only baseline, and a UAV photogrammetry-based point-cloud approach. In addition, the framework generated high-resolution plant height maps for field-scale analysis of spatial growth variability. These results underscore the importance of jointly modeling spatial and temporal characteristics for continuously evolving crop traits. The proposed framework offers an accurate, scalable, and non-destructive solution for large-scale crop phenotyping and precision agriculture.","准确的株高估算对于监测作物生长和支持精准农业管理具有重要意义。人工测量劳动强度大，而基于激光雷达（LiDAR）的方法成本高昂且需要复杂的处理。无人机摄影测量提供了一种成本较低的替代方案，但在淹水稻田中仍面临挑战，原因在于冠层变形和地形提取困难。本研究提出了Rice-STNet，一种时空深度学习框架，用于利用多时相无人机RGB影像进行端到端水稻株高估算。Rice-STNet集成了用于空间特征提取的卷积神经网络、用于时间编码的Time2Vec，以及用于建模观测日期之间时间依赖关系的门控循环单元网络。该框架利用两个生长季从稻田采集的田间数据进行了评估。Rice-STNet取得了R²为0.97、均方根误差为1.97 cm、平均绝对误差为1.14 cm的结果。其性能优于随机森林、支持向量回归、仅使用CNN的基线方法以及基于无人机摄影测量的点云方法。此外，该框架生成了高分辨率株高图，用于田块尺度空间生长变异性分析。这些结果凸显了联合建模空间与时间特征对于持续变化的作物性状的重要性。所提出的框架为大规模作物表型分析和精准农业提供了一种准确、可扩展且非破坏性的解决方案。","Agriculture","2026-09-21T00:00:00Z",{"impact":17,"substance":183,"depth":17,"authority":184,"freshness":185,"relevant":21,"comment":468},"提出时空深度学习框架Rice-STNet，用多时相无人机RGB影像实现水稻株高高精度估算，方法新颖、数据跨两个生长季，对作物表型与精准农业有实用价值。",[470],{"name":465,"url":462},[26,27,28,245,472],"作物表型",{"code":39,"slug":40,"name":28,"path":41,"parent_code":42},[475,476],{"code":32,"slug":33,"name":34,"path":35,"parent_code":36},{"code":251,"slug":252,"name":253,"path":254,"parent_code":36},{"status":44,"locked":14,"version":45,"method":46,"evidence":478},{"candidates":479,"agricultural_evidence":51},[480,484,488,493],{"code":39,"score":17,"supported":51,"title":481,"abstract":482,"tags":483},[40],[28,338,40],[28],{"code":32,"score":199,"supported":51,"title":485,"abstract":486,"tags":487},[53],[56,53],[27],{"code":251,"score":109,"supported":51,"title":489,"abstract":490,"tags":492},[],[472,491],"crop phenotyping",[472,245],{"code":64,"score":345,"supported":51,"title":494,"abstract":495,"tags":498},[],[496,497],"精准农业","precision agriculture",[26],[500,501],"无人机 RGB 水稻株高","Rice-STNet 水稻表型","无人机RGB水稻株高-3164","10.3390\u002Fagriculture16182034",{"doi":503,"openalex_id":505,"authors":506,"venue":465,"cited_by_count":72,"oa_url":462,"card":515,"direction":312,"ingested_from":92},"W7213887432",[507,510,512],{"name":508,"orcid":509},"Weiguo Wang","https:\u002F\u002Forcid.org\u002F0009-0003-4028-9363",{"name":511,"orcid":9},"Noboru Noguchi",{"name":513,"orcid":514},"Liangliang Yang","https:\u002F\u002Forcid.org\u002F0000-0002-5055-3987",{"tldr":516,"method":517,"finding":518,"direction":312,"opportunity":519},"提出Rice-STNet时空深度学习框架，用多时相无人机RGB影像估算水稻株高。","CNN提取空间特征，Time2Vec编码时间，GRU建模时序依赖，两季稻田数据验","R²达0.97、RMSE 1.97cm，优于随机森林、SVR、纯CNN及点云方法。","可迁移至其他作物与多源遥感融合，探索轻量化模型及实时田间部署。","2026-09-22T23:30:18.545958Z"]