[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-3048":3,"related-3048":75},{"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":59,"slug":62,"view_count":63,"doi":8,"paper":64,"created_at":74},3048,"基于双路径注意力与多尺度融合的作物病虫害识别网络DPMFNet","https:\u002F\u002Fwww.mdpi.com\u002F1099-4300\u002F28\u002F9\u002F1032","盐城工学院Hong Zhang、Fagen Song等联合江苏开放大学提出DPMFNet轻量级双路径网络，集成空间-通道双注意力（SCDA）与多尺度深度可分离卷积（MDSC）模块，构建AttMDSCBlock残差结构。在PlantVillage与AI Challenger 2018数据集上DPMFNet仅14.24M参数和2.55G FLOPs，跨注意力机制融合局部细节与全局上下文，轻量化金字塔策略自适应整合多分辨率特征，在复杂农田场景下兼顾精度与可部署性，为嵌入式田间设备提供高性价比方案。",null,"MDPI Entropy 28(9):1032","2026-09-20T00:00:00Z","论文",10,false,79,{"impact":16,"substance":17,"depth":18,"authority":19,"freshness":12,"relevant":20,"comment":21},16,22,18,13,1,"轻量级双路径注意力网络在两大公开数据集上兼顾精度与可部署性，对嵌入式田间设备落地有实质参考价值。",[23],{"name":9,"url":6},[25,26,27,28,29],"智慧农业","农业人工智能","病虫害识别","作物监测","轻量化模型",{"code":31,"slug":32,"name":33,"path":34,"parent_code":35},"I2","agricultural-ai","农业AI与大模型","\u002Ftopic\u002Fdigital-agriculture\u002Fagricultural-ai","I",[],{"status":38,"locked":13,"version":39,"method":40,"evidence":41},"classified","2026-10-22.2","rules",{"candidates":42,"agricultural_evidence":45},[43,49,54],{"code":31,"score":44,"supported":45,"title":46,"abstract":47,"tags":48},11,true,[27],[],[26,27,29],{"code":50,"score":20,"supported":13,"title":51,"abstract":52,"tags":53},"I1",[],[],[25],{"code":55,"score":20,"supported":13,"title":56,"abstract":57,"tags":58},"I3",[],[],[28],[60,61],"盐城工学院 DPMFNet 病虫害识别","PlantVillage AI Challenger 作物病害","盐城工学院DPMFNet病虫害识别-3048",6,{"doi":8,"openalex_id":8,"authors":65,"venue":8,"cited_by_count":66,"oa_url":8,"card":67,"direction":71,"ingested_from":73},[],0,{"tldr":68,"method":69,"finding":70,"direction":71,"opportunity":72},"提出轻量级双路径网络DPMFNet，实现复杂农田场景下的作物病虫害高精度识别。","空间-通道双注意力与多尺度深度可分离卷积，构建AttMDSCBlock残差结构。","仅14.24M参数、2.55G FLOPs，在PlantVillage与AI Challenger ","农业人工智能与决策模型","可探索真实田间多病虫害并发与边缘设备实时推理的轻量化自适应识别研究。","agent","2026-09-21T00:04:39.305757Z",{"total":63,"page":20,"page_size":63,"items":76},[77,184,269,334,403,460],{"id":78,"title":79,"url":80,"summary":81,"summary_zh":82,"content":8,"source_name":83,"source_url":80,"published_at":84,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":85,"score_detail":86,"sources":91,"tags":93,"primary_topic":95,"secondary_topics":100,"classification":102,"search_phrases":160,"slug":163,"view_count":89,"doi":164,"paper":165,"created_at":183},1398,"AI-Enabled IoT for Precision Agriculture and Smart Crop Management","https:\u002F\u002Fdoi.org\u002F10.38124\u002Fijisrt\u002F26aug1195","Agriculture is increasingly dependent on technologies that can improve productivity while reducing the consumption of water, fertilizers, pesticides, herbicides, labour and other resources. The five source chapters reviewed for this paper collectively describe the role of Artificial Intelligence (AI), Machine Learning (ML), Internet of Things (IoT), computer vision, cloud computing and embedded systems in smart farming. The summarized work focuses on continuous sensing of soil and environmental parameters, automated irrigation, crop and weed monitoring, disease identification, yield prediction and decision support. The sources identify Artificial Neural Networks (ANNs), Deep Learning, Support Vector Machines (SVMs) and Convolutional Neural Networks (CNNs) as important approaches. A prototype architecture is also described using Arduino Mega 2560, Raspberry Pi, multiple sensors, Firebase and an Apache web server. In the disease-detection prototype, 295 leaf images were divided into training, validation and testing groups, and a CNN-based mobile application produced a reported confidence score of 0.97 for an example operation with an execution time of approximately 0.88 seconds. Overall, the reviewed material indicates that integrating IoT sensing with AI\u002FML can support real-time monitoring, resource optimization and faster agricultural decisions. However, Internet dependence, cybersecurity, system complexity, adoption cost, limited datasets and reduced accuracy for visually similar crop diseases remain important challenges.","农业日益依赖于能够在提高生产力的同时减少水、肥料、农药、除草剂、劳动力及其他资源消耗的技术。本文所综述的五篇源文献共同阐述了人工智能（AI）、机器学习（ML）、物联网（IoT）、计算机视觉、云计算及嵌入式系统在智慧农业中的作用。这些研究工作的重点集中于土壤与环境参数的连续感知、自动化灌溉、作物与杂草监测、病害识别、产量预测及决策支持。相关文献将人工神经网络（ANNs）、深度学习、支持向量机（SVMs）和卷积神经网络（CNNs）视为重要技术方法。此外，文中还描述了一种基于Arduino Mega 2560、Raspberry Pi、多种传感器、Firebase及Apache Web服务器的原型架构。在病害检测原型中，295张叶片图像被划分为训练集、验证集和测试集，基于CNN的移动应用在示例操作中报告的置信度得分为0.97，执行时间约为0.88秒。总体而言，所综述的材料表明，将物联网感知与人工智能\u002F机器学习相结合，可支持实时监测、资源优化及更快速的农业决策。然而，互联网依赖、网络安全、系统复杂性、采用成本、数据集有限以及视觉上相似的作物病害识别精度降低等问题仍是重要挑战。","International Journal of Innovative Science and Research Technology (IJISRT)","2026-09-02T00:00:00Z",56,{"impact":87,"substance":18,"depth":16,"authority":88,"freshness":89,"relevant":20,"comment":90},12,8,2,"综述AI与IoT在精准农业中的应用，含原型与数据，但来源一般且时效性低。",[92],{"name":83,"url":80},[25,26,94,27,28],"物联网",{"code":96,"slug":97,"name":98,"path":99,"parent_code":35},"I5","agricultural-iot","物联网传感与边缘计算","\u002Ftopic\u002Fdigital-agriculture\u002Fagricultural-iot",[101],{"code":31,"slug":32,"name":33,"path":34,"parent_code":35},{"status":38,"locked":13,"version":39,"method":40,"evidence":103},{"candidates":104,"agricultural_evidence":45},[105,111,116,128,135,142,149,156],{"code":96,"score":18,"supported":45,"title":106,"abstract":108,"tags":110},[107],"IoT",[94,109,107],"internet of things",[94],{"code":50,"score":87,"supported":45,"title":112,"abstract":114,"tags":115},[113],"precision agriculture",[25],[25],{"code":31,"score":44,"supported":45,"title":117,"abstract":118,"tags":127},[],[119,120,121,122,123,124,125,126],"深度学习","机器学习","计算机视觉","machine learning","deep learning","artificial intelligence","computer vision","decision support",[26,27],{"code":129,"score":63,"supported":45,"title":130,"abstract":131,"tags":134},"F4",[],[132,133],"农药","除草",[],{"code":136,"score":63,"supported":45,"title":137,"abstract":138,"tags":141},"H2",[],[139,140],"土壤","soil",[],{"code":143,"score":63,"supported":45,"title":144,"abstract":145,"tags":148},"H3",[],[146,147],"灌溉","irrigation",[],{"code":150,"score":151,"supported":45,"title":152,"abstract":153,"tags":155},"F3",3,[],[154],"肥料",[],{"code":55,"score":20,"supported":13,"title":157,"abstract":158,"tags":159},[],[],[28],[161,162],"农业人工智能 病虫害识别 作物监测 智慧农业","农业人工智能 病虫害识别","农业人工智能病虫害识别作物监测智慧农业-1398","10.38124\u002Fijisrt\u002F26aug1195",{"doi":164,"openalex_id":166,"authors":167,"venue":83,"cited_by_count":66,"oa_url":80,"card":176,"direction":180,"ingested_from":182},"W7204992523",[168,170,172,174],{"name":169,"orcid":8},"Tarun Badiwal",{"name":171,"orcid":8},"Manish Jain",{"name":173,"orcid":8},"Sandeep Jayswal",{"name":175,"orcid":8},"Suresh Meena",{"tldr":177,"method":178,"finding":179,"direction":180,"opportunity":181},"综述AI与IoT在精准农业中的应用，涵盖传感、灌溉、病害检测等，并给出原型架构。","综述多篇文献，结合ANN、DL、SVM、CNN，并描述基于Arduino和树莓派","AI与IoT集成可支持实时监测和资源优化，但存在网络依赖、成本高、数据有限等挑战。","智慧农业 \u002F 农业物联网","针对视觉相似作物病害识别精度不足，可研究多模态数据融合或轻量化模型以提升鲁棒性。","openalex","2026-09-02T23:30:09.623547Z",{"id":185,"title":186,"url":187,"summary":188,"summary_zh":189,"content":8,"source_name":190,"source_url":187,"published_at":191,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":192,"score_detail":193,"sources":198,"tags":200,"primary_topic":202,"secondary_topics":208,"classification":210,"search_phrases":229,"slug":232,"view_count":20,"doi":233,"paper":234,"created_at":268},5150,"MSFANet: a multi-scale feature aggregation network for tobacco pest recognition","https:\u002F\u002Fdoi.org\u002F10.3389\u002Ffpls.2026.1871587","Introduction Tobacco is a globally significant economic crop, and pest infestation represents one of the primary factors contributing to declines in both yield and quality. However, tobacco pests encompass a wide variety of species that differ substantially in morphology and scale. Field-captured images are frequently subject to complex background interference and class imbalance, presenting considerable challenges for automated recognition. Existing deep learning approaches remain limited in their capacity for multi-scale feature extraction and discriminative feature representation under such conditions. To address these challenges, this study proposes a Multi-Scale Feature Aggregation Network (MSFANet) for tobacco pest recognition, formulated here as a single-label image classification task. Methods To address scale variation and complex field backgrounds, two task-oriented modules were developed. The Multi-Scale Pyramid Pooling (MSPP) module aggregates contextual information across multiple receptive fields through parallel feature extraction and adaptive cross-branch fusion. The Hierarchical Feature Fusion Block (HFFB) subsequently refines the feature representations through spatial gating and partial convolution. An auxiliary classification head and Focal Loss were incorporated during training to provide intermediate supervision and address class imbalance, respectively. Results The effectiveness of MSFANet was evaluated on a dataset comprising 13 tobacco-associated pest and natural-enemy categories. The model achieved accuracy, precision, recall, and F1-score of 97.63%, 97.89%, 97.59%, and 97.71%, respectively, with a parameter count of 1.21 M. Among the models trained from random initialization, MSFANet shared the highest accuracy while using the fewest parameters. Discussion By combining multi-scale contextual aggregation with hierarchical feature refinement, MSFANet achieved accurate classification across the 13 tobacco-associated categories under complex field conditions. The proposed method provides effective technical support for automated and precision-oriented pest monitoring in tobacco fields, offering practical value for advancing precise pest management and the broader development of smart agriculture.","引言 烟草是全球重要的经济作物，虫害侵染是导致其产量和品质下降的主要因素之一。然而，烟草害虫种类繁多，形态和尺度差异显著。田间采集的图像常受到复杂背景干扰和类别不平衡的影响，给自动化识别带来了相当大的挑战。现有深度学习方法在此类条件下进行多尺度特征提取和判别性特征表示的能力仍然有限。为应对这些挑战，本研究提出了一种用于烟草害虫识别的多尺度特征聚合网络（Multi-Scale Feature Aggregation Network，MSFANet），并将其形式化为单标签图像分类任务。方法 针对尺度变化和复杂田间背景，开发了两个面向任务的模块。多尺度金字塔池化（Multi-Scale Pyramid Pooling，MSPP）模块通过并行特征提取和自适应跨分支融合，聚合多个感受野的上下文信息。分层特征融合模块（Hierarchical Feature Fusion Block，HFFB）随后通过空间门控和部分卷积对特征表示进行细化。训练过程中引入辅助分类头和Focal Loss，分别提供中间监督和处理类别不平衡。结果 在包含13个烟草相关害虫及天敌类别的数据集上评估了MSFANet的有效性。模型的准确率、精确率、召回率和F1分数分别为97.63%、97.89%、97.59%和97.71%，参数量为1.21 M。在随机初始化训练的模型中，MSFANet以最少的参数量取得了最高准确率。讨论 通过结合多尺度上下文聚合与分层特征细化，MSFANet在复杂田间条件下实现了对13个烟草相关类别的准确分类。所提方法为烟草田间害虫的自动化和精准监测提供了有效的技术支持，对推进精准害虫管理和智慧农业的更广泛发展具有实用价值。","Frontiers in Plant Science","2026-10-16T00:00:00Z",76,{"impact":16,"substance":194,"depth":195,"authority":19,"freshness":196,"relevant":20,"comment":197},21,17,9,"该论文提出轻量化多尺度特征聚合网络，在烟草害虫识别上取得97%以上精度且参数量仅1.21M，方法新颖、数据扎实，对智慧农业虫情监测有实用价值，值得进入每日精选。",[199],{"name":190,"url":187},[25,26,119,27,201],"烟草",{"code":203,"slug":204,"name":205,"path":206,"parent_code":207},"A6","cash-crops","棉麻糖与特色经作","\u002Ftopic\u002Ffield-crops\u002Fcash-crops","A",[209],{"code":31,"slug":32,"name":33,"path":34,"parent_code":35},{"status":38,"locked":13,"version":39,"method":40,"evidence":211},{"candidates":212,"agricultural_evidence":45},[213,219,223],{"code":203,"score":214,"supported":45,"title":215,"abstract":217,"tags":218},15,[216],"tobacco",[201,216],[201],{"code":31,"score":196,"supported":45,"title":220,"abstract":221,"tags":222},[],[119,123],[26,119,27],{"code":50,"score":224,"supported":45,"title":225,"abstract":226,"tags":228},7,[],[25,227],"smart agriculture",[25],[230,231],"MSFANet 烟草害虫识别","多尺度特征聚合 害虫识别","MSFANet烟草害虫识别-5150","10.3389\u002Ffpls.2026.1871587",{"doi":233,"openalex_id":235,"authors":236,"venue":190,"cited_by_count":66,"oa_url":262,"card":263,"direction":180,"ingested_from":182},"W7222603139",[237,239,241,243,245,247,249,251,254,257,260],{"name":238,"orcid":8},"Shun He",{"name":240,"orcid":8},"Sifan Li",{"name":242,"orcid":8},"Xinjian Li",{"name":244,"orcid":8},"Qianqian Luo",{"name":246,"orcid":8},"Dong Sufen",{"name":248,"orcid":8},"Lu Lei",{"name":250,"orcid":8},"Xingjiang Chen",{"name":252,"orcid":253},"Liuti Cai","https:\u002F\u002Forcid.org\u002F0000-0002-9151-8002",{"name":255,"orcid":256},"Jian‐Yu Meng","https:\u002F\u002Forcid.org\u002F0000-0002-5796-1086",{"name":258,"orcid":259},"Ning Lu","https:\u002F\u002Forcid.org\u002F0000-0003-0214-5038",{"name":261,"orcid":8},"Hancheng Wang","https:\u002F\u002Fwww.frontiersin.org\u002Fjournals\u002Fplant-science\u002Farticles\u002F10.3389\u002Ffpls.2026.1871587\u002Fpdf",{"tldr":264,"method":265,"finding":266,"direction":71,"opportunity":267},"提出MSFANet多尺度特征聚合网络，实现复杂田间背景下烟草害虫的自动识别。","多尺度金字塔池化MSPP、分层特征融合HFFB、辅助分类头与Focal Loss","模型准确率97.63%、参数量仅1.21M，在随机初始化模型中精度最高且参数最少。","可探索轻量多尺度网络在边缘设备部署及跨作物害虫迁移识别，结合类不平衡与复杂背景的鲁棒性研究。","2026-10-17T23:30:08.695455Z",{"id":270,"title":271,"url":272,"summary":273,"summary_zh":274,"content":8,"source_name":275,"source_url":272,"published_at":276,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":277,"score_detail":278,"sources":281,"tags":283,"primary_topic":286,"secondary_topics":287,"classification":288,"search_phrases":311,"slug":314,"view_count":89,"doi":315,"paper":316,"created_at":333},4990,"Smart Crop Advisory System","https:\u002F\u002Fdoi.org\u002F10.55041\u002Fijcope.v2i10.081","The Smart Crop Advisory System is a web-based agricultural decision-support platform that uses Artificial Intelligence (AI) and Machine Learning (ML) technologies to assist farmers in making informed farming decisions. Traditional agriculture faces several challenges, including unsuitable crop selection, unpredictable weather conditions, crop diseases, and fluctuations in market prices. The proposed system integrates essential agricultural services, including crop recommendation, weather forecasting, crop disease detection, agricultural market price information, and an AI-based multilingual voice assistant. It also provides farmer profile management, crop price trend analysis, and rainfall trend visualization through an interactive dashboard. The crop recommendation module helps farmers identify suitable crops based on relevant agricultural parameters, while the disease detection module supports the identification of potential crop diseases. Weather and market information help farmers plan agricultural activities and make informed selling decisions. The multilingual voice assistant aims to improve accessibility by enabling farmers to interact with the system through voice-based communication. Overall, the Smart Crop Advisory System aims to improve access to agricultural information, support effective decision-making, promote efficient resource utilization, and encourage sustainable farming practices. Keywords: Smart Agriculture, Crop Recommendation, Machine Learning, Crop Disease Detection, Weather Forecasting, Market Price Analysis, Voice Assistant, Sustainable Agriculture.","智能作物咨询系统是一个基于网络的农业决策支持平台，利用人工智能（AI）和机器学习（ML）技术帮助农民做出明智的农作决策。传统农业面临诸多挑战，包括作物选择不当、天气条件不可预测、作物病害以及市场价格波动。所提出的系统整合了基本农业服务，包括作物推荐、天气预报、作物病害检测、农产品市场价格信息以及基于AI的多语言语音助手。它还通过交互式仪表板提供农民档案管理、作物价格趋势分析和降雨趋势可视化。作物推荐模块帮助农民根据相关农业参数识别适宜作物，而病害检测模块支持识别潜在作物病害。天气和市场信息帮助农民规划农业活动并做出明智的销售决策。多语言语音助手旨在通过使农民能够以语音方式与系统交互来提高可及性。总体而言，智能作物咨询系统旨在改善农业信息的获取，支持有效决策，促进资源高效利用，并鼓励可持续农业实践。关键词：智慧农业、作物推荐、机器学习、作物病害检测、天气预报、市场价格分析、语音助手、可持续农业。","International Journal of Creative and Open Research in Engineering and Management","2026-10-14T00:00:00Z",48,{"impact":88,"substance":279,"depth":87,"authority":63,"freshness":88,"relevant":20,"comment":280},14,"面向农户的AI农业决策支持系统论文，功能集成度较高但方法新颖性与数据规模有限，属细分领域进展，可作为智慧农业主题聚合素材。",[282],{"name":275,"url":272},[25,26,27,284,285],"农业气象","作物推荐",{"code":31,"slug":32,"name":33,"path":34,"parent_code":35},[],{"status":38,"locked":13,"version":39,"method":40,"evidence":289},{"candidates":290,"agricultural_evidence":45},[291,295,299,306],{"code":31,"score":44,"supported":45,"title":292,"abstract":293,"tags":294},[],[120,122,124],[26,27],{"code":50,"score":224,"supported":45,"title":296,"abstract":297,"tags":298},[],[25,227],[25],{"code":300,"score":63,"supported":45,"title":301,"abstract":302,"tags":305},"H8",[],[303,304],"可持续农业","sustainable agriculture",[],{"code":307,"score":20,"supported":13,"title":308,"abstract":309,"tags":310},"H7",[],[],[284],[312,313],"AI 作物推荐 语音助手","农业人工智能 病虫害识别 作物推荐 农业气象","AI作物推荐语音助手-4990","10.55041\u002Fijcope.v2i10.081",{"doi":315,"openalex_id":317,"authors":318,"venue":275,"cited_by_count":66,"oa_url":327,"card":328,"direction":180,"ingested_from":182},"W7221803029",[319,321,323,325],{"name":320,"orcid":8},"Bhuvana H P",{"name":322,"orcid":8},"Kavya C M",{"name":324,"orcid":8},"Manjula S",{"name":326,"orcid":8},"Priyanka A","https:\u002F\u002Fijcope.org\u002Fwp-content\u002Fuploads\u002F2026\u002F10\u002FSmart-Crop-Advisory-System.pdf",{"tldr":329,"method":330,"finding":331,"direction":71,"opportunity":332},"构建基于AI\u002FML的Web农业决策支持平台，为农民提供作物推荐、病害检测、天气与市场价格等服务。","集成机器学习模型、多语言语音助手与交互式仪表盘，整合天气、市场及农业参数数据。","系统可提升农业信息获取与决策效率，促进资源高效利用和可持续耕作。","可探索多源数据融合的轻量化模型与本地化语音交互，提升小农户采纳率与实时决策精度。","2026-10-15T23:30:09.048921Z",{"id":335,"title":336,"url":337,"summary":338,"summary_zh":339,"content":8,"source_name":340,"source_url":337,"published_at":341,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":342,"score_detail":343,"sources":346,"tags":348,"primary_topic":351,"secondary_topics":356,"classification":358,"search_phrases":374,"slug":377,"view_count":66,"doi":378,"paper":379,"created_at":402},4818,"Hybrid Deep Feature Extraction and Ensemble Machine Learning for Potato Leaf Disease Classification in Uncontrolled Field Conditions","https:\u002F\u002Fdoi.org\u002F10.15832\u002Fankutbd.1837896","Early and accurate diagnosis of potato leaf diseases is critical for preventing yield losses under real field conditions. This study proposes a hybrid deep feature–based machine learning framework that combines transfer learning and ensemble classifiers to automatically identify seven potato leaf disease classes from images. A total of 4,260 images from the publicly available Potato Leaf Disease Dataset were organized into seven categories: Bacteria, Virus, Pest, Fungi, Nematode, Phytophthora, and Healthy. Real-time image augmentation was performed using Keras ImageDataGenerator to improve strength to field variability. Deep feature vectors were extracted from three pre-trained convolutional neural networks (ResNet50, EfficientNetB0, and MobileNetV2), each followed by a lightweight dense head, and concatenated into 768-dimensional feature representations. After standardization, dimensionality reduction was applied via principal component analysis, and class imbalance in the training set was addressed using the SMOTE-ENN hybrid resampling method. On these deep features, 32 classical machine learning and ensemble models including SVMs, tree-based methods, boosting algorithms, multilayer perceptrons, and four ensemble schemes, were trained. Model performance was assessed by three-fold cross-validation on the training set and final evaluation on the held-out validation set using accuracy, precision, recall, and macro F1-score. Ensemble models outperformed individual classifiers. The best performance was obtained with the StackingClassifier_L2 model, which combines XGBoost, LightGBM, Random Forest, SVM with RBF kernel, and Extra Trees within a stacking framework. This model achieved an accuracy of 81.34% and a macro F1-score of 81.74% on the validation set. The proposed hybrid system, which integrates multi-backbone deep feature extraction with comprehensive classical and ensemble classifiers, provides a robust and practical solution for automatic potato leaf disease classification in uncontrolled field conditions. The results suggest that stacking-based ensembles on top of deep features can capture complex visual patterns more effectively than single models and may support decision-making in precision agriculture.","在马铃薯实际田间条件下，早期准确地诊断叶片病害对于防止产量损失至关重要。本研究提出了一种基于混合深度特征的机器学习框架，该框架结合迁移学习和集成分类器，从图像中自动识别七类马铃薯叶片病害。来自公开可用的马铃薯叶片病害数据集的4,260张图像被分为七个类别：细菌（Bacteria）、病毒（Virus）、害虫（Pest）、真菌（Fungi）、线虫（Nematode）、疫霉（Phytophthora）和健康（Healthy）。使用Keras ImageDataGenerator进行实时图像增强，以提高对田间变异的鲁棒性。从三个预训练的卷积神经网络（ResNet50、EfficientNetB0和MobileNetV2）中提取深度特征向量，每个网络后接一个轻量级密集头，并将其拼接为768维特征表示。经过标准化后，通过主成分分析进行降维，并使用SMOTE-ENN混合重采样方法解决训练集中的类别不平衡问题。在这些深度特征上，训练了32种经典机器学习和集成模型，包括支持向量机、基于树的方法、提升算法、多层感知机以及四种集成方案。通过训练集上的三折交叉验证和保留验证集上的最终评估，使用准确率、精确率、召回率和宏F1分数评估模型性能。集成模型优于单个分类器。最佳性能由StackingClassifier_L2模型获得，该模型在堆叠框架内结合了XGBoost、LightGBM、随机森林、带RBF核的支持向量机和极端随机树。该模型在验证集上达到了81.34%的准确率和81.74%的宏F1分数。所提出的混合系统将多骨干深度特征提取与全面的经典分类器和集成分类器相结合，为无控制田间条件下的马铃薯叶片病害自动分类提供了一种稳健且实用的解决方案。结果表明，基于深度特征的堆叠集成能够比单一模型更有效地捕捉复杂的视觉模式，并可能支持精准农业中的决策。","Tarım Bilimleri Dergisi","2026-10-12T00:00:00Z",65,{"impact":88,"substance":344,"depth":16,"authority":19,"freshness":88,"relevant":20,"comment":345},20,"方法组合较全面、数据规模明确，但验证集准确率仅81.34%，属细分领域技术进展，可作主题聚合素材而非头条精选。",[347],{"name":340,"url":337},[25,26,349,27,350],"马铃薯","精准农业",{"code":352,"slug":353,"name":354,"path":355,"parent_code":207},"A5","tubers-grains","薯类与杂粮","\u002Ftopic\u002Ffield-crops\u002Ftubers-grains",[357],{"code":31,"slug":32,"name":33,"path":34,"parent_code":35},{"status":38,"locked":13,"version":39,"method":40,"evidence":359},{"candidates":360,"agricultural_evidence":45},[361,365,370],{"code":31,"score":16,"supported":45,"title":362,"abstract":363,"tags":364},[122],[120,122],[26,27],{"code":352,"score":214,"supported":45,"title":366,"abstract":368,"tags":369},[367],"potato",[349,367],[349],{"code":50,"score":88,"supported":45,"title":371,"abstract":372,"tags":373},[],[350,113],[25,350],[375,376],"马铃薯 叶片病害 深度学习","Potato Leaf Disease Dataset 集成学习","马铃薯叶片病害深度学习-4818","10.15832\u002Fankutbd.1837896",{"doi":378,"openalex_id":380,"authors":381,"venue":340,"cited_by_count":66,"oa_url":396,"card":397,"direction":180,"ingested_from":182},"W7221546496",[382,385,387,390,393],{"name":383,"orcid":384},"Cihan Ünal","https:\u002F\u002Forcid.org\u002F0000-0002-5255-4078",{"name":386,"orcid":8},"Cengizhan Tumer",{"name":388,"orcid":389},"Elham Tahsin Yasin","https:\u002F\u002Forcid.org\u002F0000-0003-3246-6000",{"name":391,"orcid":392},"İlkay Çınar","https:\u002F\u002Forcid.org\u002F0000-0003-0611-3316",{"name":394,"orcid":395},"Murat Köklü","https:\u002F\u002Forcid.org\u002F0000-0002-2737-2360","https:\u002F\u002Fdergipark.org.tr\u002Fen\u002Fdownload\u002Farticle-file\u002F5487015",{"tldr":398,"method":399,"finding":400,"direction":71,"opportunity":401},"融合三种CNN深度特征与集成学习，实现田间马铃薯叶片七类病害分类。","ResNet50\u002FEfficientNetB0\u002FMobileNetV2特征拼接+","StackingClassifier_L2最优，验证集准确率81.34%、宏F1为81.74%。","可探索轻量化模型与多模态数据融合，提升非受控田间小样本病害识别鲁棒性。","2026-10-13T23:30:16.021066Z",{"id":404,"title":405,"url":406,"summary":407,"summary_zh":408,"content":8,"source_name":340,"source_url":406,"published_at":341,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":409,"score_detail":410,"sources":412,"tags":414,"primary_topic":417,"secondary_topics":421,"classification":423,"search_phrases":439,"slug":442,"view_count":66,"doi":443,"paper":444,"created_at":459},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。",80,{"impact":18,"substance":17,"depth":18,"authority":19,"freshness":196,"relevant":20,"comment":411},"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。",[413],{"name":340,"url":406},[25,26,415,27,416],"水稻","联邦学习",{"code":418,"slug":419,"name":415,"path":420,"parent_code":207},"A1","rice","\u002Ftopic\u002Ffield-crops\u002Frice",[422],{"code":31,"slug":32,"name":33,"path":34,"parent_code":35},{"status":38,"locked":13,"version":39,"method":40,"evidence":424},{"candidates":425,"agricultural_evidence":45},[426,431,435],{"code":418,"score":214,"supported":45,"title":427,"abstract":429,"tags":430},[428],"paddy",[415,428],[415],{"code":31,"score":87,"supported":45,"title":432,"abstract":433,"tags":434},[],[120,122,126],[26,27,416],{"code":50,"score":20,"supported":13,"title":436,"abstract":437,"tags":438},[],[],[25],[440,441],"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":443,"openalex_id":445,"authors":446,"venue":340,"cited_by_count":66,"oa_url":453,"card":454,"direction":180,"ingested_from":182},"W7221522859",[447,450],{"name":448,"orcid":449},"B Johnson","https:\u002F\u002Forcid.org\u002F0009-0009-8281-4051",{"name":451,"orcid":452},"Chandrakumar Thangavel","https:\u002F\u002Forcid.org\u002F0000-0002-1186-5988","https:\u002F\u002Fdergipark.org.tr\u002Fen\u002Fdownload\u002Farticle-file\u002F5486058",{"tldr":455,"method":456,"finding":457,"direction":71,"opportunity":458},"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":461,"title":462,"url":463,"summary":464,"summary_zh":465,"content":8,"source_name":466,"source_url":463,"published_at":467,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":468,"score_detail":469,"sources":471,"tags":473,"primary_topic":476,"secondary_topics":477,"classification":478,"search_phrases":504,"slug":507,"view_count":508,"doi":509,"paper":510,"created_at":521},4798,"Adaptability of Lightweight AI Models for Embedded Deployment: A Case Study of an Agricultural Monitoring Robot","https:\u002F\u002Fdoi.org\u002F10.54254\u002F2755-2721\u002F2026.ld37676","Many artificial intelligence programs require powerful computing hardware, whereas agricultural robots, drones, and smart cameras rely on small embedded devices with limited computing power, battery capacity, and heat dissipation. Lightweight models reduce parameters and computational costs while maintaining acceptable accuracy, enabling AI deployment on resource-constraint network devices. This paper reviews key lightweighting methods, including compact network architectures, model pruning, model quantization, and knowledge distillation. It compares common platforms such as Jetson, Raspberry Pi, and microcontrollers. Using an agricultural monitoring robot as an example, this study proposes a system combining a camera, a lightweight YOLO model, Jetson, and a microcontroller for real-time crop monitoring. Experiments are designed to evaluate accuracy, speed, power consumption, temperature, and battery life. Rather than pursuing maximum model complexity, this paper focuses on explaining how to select a model that is truly suitable for the actual device. For agricultural robots, effective deployment requires sufficient accuracy, real-time performance, and low energy consumption.","许多人工智能程序需要强大的计算硬件，而农业机器人、无人机和智能摄像头则依赖于计算能力、电池容量和散热能力均有限的小型嵌入式设备。轻量级模型在保持可接受精度的同时减少参数和计算成本，使人工智能能够部署在资源受限的网络设备上。本文综述了关键的轻量化方法，包括紧凑网络架构、模型剪枝、模型量化和知识蒸馏。本文比较了Jetson、树莓派和微控制器等常见平台。以农业监测机器人为例，本研究提出了一种结合摄像头、轻量级YOLO模型、Jetson和微控制器的系统，用于实时作物监测。实验旨在评估精度、速度、功耗、温度和电池寿命。本文并非追求最大化的模型复杂度，而是着重解释如何选择真正适合实际设备的模型。对于农业机器人而言，有效部署需要足够的精度、实时性能和低能耗。","Applied and Computational Engineering","2026-10-13T00:00:00Z",69,{"impact":87,"substance":18,"depth":16,"authority":19,"freshness":12,"relevant":20,"comment":470},"系统梳理轻量化模型在农业嵌入式设备上的部署路径，并以监测机器人案例给出可复现方案，对智慧农业工程落地有参考价值。",[472],{"name":466,"url":463},[25,26,474,475,28],"农业机器人","边缘计算",{"code":31,"slug":32,"name":33,"path":34,"parent_code":35},[],{"status":38,"locked":13,"version":39,"method":40,"evidence":479},{"candidates":480,"agricultural_evidence":45},[481,486,492,496,500],{"code":31,"score":224,"supported":45,"title":482,"abstract":483,"tags":485},[],[124,484],"YOLO",[26],{"code":487,"score":488,"supported":45,"title":489,"abstract":490,"tags":491},"F6",4,[],[474],[474],{"code":50,"score":20,"supported":13,"title":493,"abstract":494,"tags":495},[],[],[25],{"code":55,"score":20,"supported":13,"title":497,"abstract":498,"tags":499},[],[],[28],{"code":96,"score":20,"supported":13,"title":501,"abstract":502,"tags":503},[],[],[475],[505,506],"农业监测机器人 轻量YOLO Jetson","嵌入式AI 模型轻量化","农业监测机器人轻量YOLOJetson-4798",5,"10.54254\u002F2755-2721\u002F2026.ld37676",{"doi":509,"openalex_id":511,"authors":512,"venue":466,"cited_by_count":66,"oa_url":515,"card":516,"direction":180,"ingested_from":182},"W7221561538",[513],{"name":514,"orcid":8},"Jingen Zhou","https:\u002F\u002Face.ewapub.com\u002Farticle\u002Fview\u002F37676.pdf",{"tldr":517,"method":518,"finding":519,"direction":180,"opportunity":520},"综述轻量化AI方法，并以农业监测机器人为例，提出相机+轻量YOLO+Jetson+单片机的嵌入式部署","综述网络压缩、剪枝、量化、蒸馏，对比Jetson、树莓派、单片机，设计农业机器人","农业机器人部署需兼顾足够精度、实时性与低能耗，而非追求最大模型复杂度。","可研究面向田间多变环境的轻量化模型自适应选择与能效-精度联合优化部署策略。","2026-10-13T23:30:08.379081Z"]