[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-1536":3,"related-1536":93},{"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":9,"secondary_topics":31,"classification":32,"search_phrases":61,"slug":64,"view_count":51,"doi":65,"paper":66,"created_at":92},1536,"A Non-Invasive Approach for Detecting Water Adulteration in Orange Juice Using Computer Vision and Deep Learning","https:\u002F\u002Fdoi.org\u002F10.1016\u002Fj.jfca.2026.109484","Food fraud due to water adulteration in orange juice is a significant concern for the beverage industry. This study aimed to develop a rapid and non-invasive computer-vision method based on deep learning to detect and classify water-adulteration levels ranging from 1% to 15% in three orange juice products representing freshly squeezed juice, juice from concentrate, and orange nectar. High-resolution images were acquired under two shutter-speed-based exposure conditions: the higher-exposure acquisition condition (1\u002F30 s) and the lower-exposure acquisition condition (1\u002F250 s), while maintaining constant scene illumination, and were analyzed using ResNet50 convolutional neural networks. In the image-level hold-out test, the model trained with images acquired at 1\u002F250 s achieved an accuracy of 88.3%, whereas the model trained at 1\u002F30 s reached 83.1%. To provide a more stringent assessment of sample-level transferability, the previously trained and fixed 1\u002F250 s model was subsequently evaluated using 240 independently prepared and blindly coded samples obtained from subsequent purchases of the same commercial products. This independent blind validation achieved a 24-class accuracy of 86.7%. Misclassifications occurred predominantly between adjacent or closely related water-adulteration levels within the same juice product. When the independent-validation predictions were collapsed into a binary pure-versus-adulterated screening task, the model achieved 100.0% sensitivity, 93.3% specificity, 99.2% accuracy, and 96.7% balanced accuracy. These findings support the feasibility of the method as a rapid pre-screening tool for detecting visible-image patterns associated with controlled water dilution and show that detecting the presence of adulteration is more reliable than assigning an exact value to closely spaced adulteration levels. The independent validation provides evidence of transferability to newly prepared samples from subsequent purchases of the same products. However, further validation across additional brands, production batches, orange origins, seasons, and acquisition environments is required before broader applicability can be established.","橙汁中因掺水导致的食品欺诈是饮料行业关注的重要问题。本研究旨在开发一种基于深度学习的快速、非侵入性计算机视觉方法，用于检测和分类三种橙汁产品（鲜榨汁、浓缩还原汁和橙汁饮料）中1%至15%的掺水水平。在保持场景照明恒定的条件下，分别以两种基于快门速度的曝光条件采集高分辨率图像：高曝光采集条件（1\u002F30秒）和低曝光采集条件（1\u002F250秒），并使用ResNet50卷积神经网络进行分析。在图像级留出测试中，以1\u002F250秒采集图像训练的模型准确率达到88.3%，而以1\u002F30秒训练的模型准确率为83.1%。为对样本级可迁移性进行更严格的评估，先前训练并固定的1\u002F250秒模型随后被用于评估240个独立制备、盲法编码的样本，这些样本来自后续购买的同款商业产品。该独立盲法验证实现了24类别86.7%的准确率。误分类主要发生在同一果汁产品内相邻或相近的掺水水平之间。当独立验证预测结果被合并为二元纯正与掺假筛查任务时，模型实现了100.0%的灵敏度、93.3%的特异度、99.2%的准确率和96.7%的平衡准确率。这些发现支持该方法作为快速预筛查工具的可行性，用于检测与受控水稀释相关的可见图像模式，并表明检测掺假的存在比精确赋值于间隔较小的掺水水平更为可靠。独立验证提供了该方法对后续购买同款产品新制备样本具有可迁移性的证据。然而，在确立更广泛的适用性之前，仍需在更多品牌、生产批次、橙子产地、季节和采集环境中进行进一步验证。",null,"Journal of Food Composition and Analysis","2026-09-01T00:00:00Z","论文",10,false,71,{"impact":17,"substance":18,"depth":19,"authority":20,"freshness":21,"relevant":22,"comment":23},12,20,18,14,7,1,"研究提出基于深度学习的橙汁掺水无损检测方法，独立验证准确率高，对食品安全监管有参考价值。",[25],{"name":10,"url":6},[27,28,29,30],"智慧农业","农业人工智能","无损检测","食品安全",[],{"status":33,"locked":14,"version":34,"method":35,"evidence":36},"needs_review","2026-10-22.2","rules",{"candidates":37,"agricultural_evidence":14,"reason":60},[38,49,55],{"code":39,"score":40,"supported":41,"title":42,"abstract":45,"tags":48},"I2",26,true,[43,44],"deep learning","computer vision",[46,47,43],"深度学习","计算机视觉",[28],{"code":50,"score":51,"supported":14,"title":52,"abstract":53,"tags":54},"E4",2,[],[],[29,30],{"code":56,"score":22,"supported":14,"title":57,"abstract":58,"tags":59},"I1",[],[],[27],"农业相关性需要复核",[62,63],"农业人工智能 无损检测 智慧农业 食品安全","农业人工智能 无损检测","农业人工智能无损检测智慧农业食品安全-1536","10.1016\u002Fj.jfca.2026.109484",{"doi":65,"openalex_id":67,"authors":68,"venue":10,"cited_by_count":83,"oa_url":84,"card":85,"direction":89,"ingested_from":91},"W7204957864",[69,71,73,76,78,80],{"name":70,"orcid":9},"Ana M. Pérez-Calabuig",{"name":72,"orcid":9},"Sandra Pradana‐López",{"name":74,"orcid":75},"John C. Cancilla","https:\u002F\u002Forcid.org\u002F0000-0003-3645-7224",{"name":77,"orcid":9},"María Luz Mena",{"name":79,"orcid":9},"Carlos López-Pingarrón",{"name":81,"orcid":82},"José S. Torrecilla","https:\u002F\u002Forcid.org\u002F0000-0003-1209-203X",0,"https:\u002F\u002Fwww.sciencedirect.com\u002Fscience\u002Farticle\u002Fpii\u002FS0889157526006277\u002Fpdf",{"tldr":86,"method":87,"finding":88,"direction":89,"opportunity":90},"用深度学习计算机视觉检测橙汁掺水，实现非侵入式快速筛查。","采集不同曝光图像，用ResNet50分类掺水等级，独立盲样验证。","掺水检测准确率高，区分精确等级较难，模型可迁移至新样本。","农业人工智能与决策模型","可扩展至其他食品掺假检测，需跨品牌、批次、季节等验证，或开发便携设备。","openalex","2026-09-03T23:30:50.596809Z",{"total":94,"page":22,"page_size":94,"items":95},6,[96,205,292,346,416,480],{"id":97,"title":98,"url":99,"summary":100,"summary_zh":101,"content":9,"source_name":102,"source_url":99,"published_at":103,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":104,"score_detail":105,"sources":109,"tags":111,"primary_topic":114,"secondary_topics":120,"classification":125,"search_phrases":178,"slug":181,"view_count":83,"doi":182,"paper":183,"created_at":204},5582,"AI-Based RGB Imaging and 14-Channel Multispectral Food Quality Scanner","https:\u002F\u002Fdoi.org\u002F10.22214\u002Fijraset.2026.85255","Food-quality assessment plays an important role in food safety, agricultural processing, storage, and supply-chain management. Conventional inspection methods depend on visual judgement or laboratory-based testing, which may require considerable time, specialised equipment, and manual effort. Recent developments in computer vision, multispectral sensing, and artificial intelligence (AI) provide opportunities for automated, non-destructive food-quality assessment. This paper reviews the application of RGB imaging, multispectral sensing, machine learning, and deep learning for fruit-ripeness estimation, surfacedefect detection, grain classification, and food-authenticity assessment. The reviewed literature demonstrates that RGB imaging can extract visible features such as colour, shape, and texture, whereas multispectral sensing provides wavelength-dependent information associated with food characteristics. Published studies have reported promising results in tomato-ripeness classification, apple-defect detection, rice-grain analysis, and honey-adulteration screening. Based on these findings, a compact food-quality scanner integrating an RGB camera, a 14-channel multispectral sensor, controlled illumination, and an embedded processing unit is proposed. The review discusses sensor integration, data preprocessing, feature fusion, classification methods, evaluation metrics, and implementation challenges. It also identifies the need for broader datasets, standardised illumination, independent validation, and resource-efficient AI models to improve the reliability of portable food-inspection systems.","食品质量评估在食品安全、农产品加工、储运和供应链管理中发挥着重要作用。传统检测方法依赖人工目视判断或实验室检测，往往需要较长时间、专用设备和人工投入。近年来，计算机视觉、多光谱传感和人工智能（AI）的发展为自动化、无损食品质量评估提供了新的机遇。本文综述了RGB成像、多光谱传感、机器学习和深度学习在果实成熟度估计、表面缺陷检测、谷物分类和食品真实性评估中的应用。所综述的文献表明，RGB成像能够提取颜色、形状和纹理等可见特征，而多光谱传感则可提供与食品特性相关的波长依赖信息。已发表研究在番茄成熟度分类、苹果缺陷检测、稻米籽粒分析和蜂蜜掺假筛查等方面取得了良好结果。基于上述研究结果，本文提出一种集成了RGB相机、14通道多光谱传感器、受控照明和嵌入式处理单元的紧凑型食品质量扫描仪。综述讨论了传感器集成、数据预处理、特征融合、分类方法、评价指标和实现挑战，并指出需要更广泛的数据集、标准化照明、独立验证和资源高效的AI模型，以提高便携式食品检测系统的可靠性。","International Journal for Research in Applied Science and Engineering Technology","2026-10-21T00:00:00Z",65,{"impact":17,"substance":19,"depth":106,"authority":107,"freshness":13,"relevant":22,"comment":108},16,9,"综述性论文，提出集成RGB与14通道多光谱的便携食品质量扫描仪方案，方法框架清晰但缺乏自有实验数据，属细分领域技术进展，可作参考但非重大突破。",[110],{"name":102,"url":99},[27,28,29,112,113],"多光谱成像","食品质量检测",{"code":115,"slug":116,"name":117,"path":118,"parent_code":119},"I3","remote-sensing","遥感与农情监测","\u002Ftopic\u002Fdigital-agriculture\u002Fremote-sensing","I",[121],{"code":39,"slug":122,"name":123,"path":124,"parent_code":119},"agricultural-ai","农业AI与大模型","\u002Ftopic\u002Fdigital-agriculture\u002Fagricultural-ai",{"status":126,"locked":14,"version":34,"method":35,"evidence":127},"classified",{"candidates":128,"agricultural_evidence":41},[129,134,141,146,153,160,167,174],{"code":115,"score":17,"supported":41,"title":130,"abstract":132,"tags":133},[131],"multispectral",[131],[112],{"code":39,"score":13,"supported":41,"title":135,"abstract":136,"tags":140},[],[46,137,47,138,43,139,44],"机器学习","machine learning","artificial intelligence",[28],{"code":50,"score":21,"supported":41,"title":142,"abstract":143,"tags":145},[],[30,144],"food safety",[29],{"code":147,"score":94,"supported":41,"title":148,"abstract":149,"tags":152},"A1",[],[150,151],"稻米","rice",[],{"code":154,"score":94,"supported":41,"title":155,"abstract":156,"tags":159},"B1",[],[157,158],"番茄","tomato",[],{"code":161,"score":94,"supported":41,"title":162,"abstract":163,"tags":166},"B2",[],[164,165],"苹果","apple",[],{"code":168,"score":169,"supported":41,"title":170,"abstract":171,"tags":173},"E1",3,[],[172],"农产品加工",[],{"code":56,"score":22,"supported":14,"title":175,"abstract":176,"tags":177},[],[],[27],[179,180],"RGB成像 多光谱 食品质量","14通道 多光谱 扫描仪","RGB成像多光谱食品质量-5582","10.22214\u002Fijraset.2026.85255",{"doi":182,"openalex_id":184,"authors":185,"venue":102,"cited_by_count":83,"oa_url":99,"card":198,"direction":203,"ingested_from":91},"W7223139905",[186,188,190,192,194,196],{"name":187,"orcid":9},"Eshwari Take",{"name":189,"orcid":9},"Avanish Raut",{"name":191,"orcid":9},"Vedant Joshi",{"name":193,"orcid":9},"Shravani Akolkar",{"name":195,"orcid":9},"R. M. Gharat",{"name":197,"orcid":9},"Shruti Dive",{"tldr":199,"method":200,"finding":201,"direction":89,"opportunity":202},"综述RGB与多光谱成像结合AI的食品品质无损检测，并提出一款集成14通道多光谱传感器的紧凑型扫描仪方","综述RGB成像、多光谱传感与机器学习\u002F深度学习在食品检测中的应用，并提出传感器集","RGB与多光谱融合在番茄成熟度、苹果缺陷、稻米分类和蜂蜜掺假检测中表现良好。","可研究多源光谱特征融合与轻量化模型，构建标准化数据集和便携设备独立验证体系。","数字乡村与农业信息化","2026-10-22T11:20:48.684398Z",{"id":206,"title":207,"url":208,"summary":209,"summary_zh":210,"content":9,"source_name":211,"source_url":208,"published_at":212,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":213,"score_detail":214,"sources":218,"tags":220,"primary_topic":222,"secondary_topics":223,"classification":229,"search_phrases":265,"slug":268,"view_count":83,"doi":269,"paper":270,"created_at":291},5400,"Artificial intelligence in food safety: a bibliometric and empirical synthesis of emerging technologies and research frontiers","https:\u002F\u002Fdoi.org\u002F10.1108\u002Fbfj-06-2025-0819","Purpose This study aimed to identify recent research trends in the integration of artificial intelligence (AI) with food safety and security through a comprehensive bibliometric analysis. Design\u002Fmethodology\u002Fapproach A total of 2,782 documents published between 2021 and 2025 were retrieved from the Web of Science Core Collection. The dataset was refined by removing duplicates, filtering document types and excluding irrelevant records. VOSviewer was then used to conduct co-authorship, keyword co-occurrence, citation, co-citation and bibliographic coupling analyses. Findings Machine learning and deep learning emerged as core technologies associated with food security, contamination detection and quality prediction. Keyword clustering demonstrated interdisciplinary connections with agriculture, climate change and public health. Citation and co-citation analyses identified three major intellectual foundations: traditional machine learning, computer vision-based deep learning and time-series modeling. China, India and the United States were the most influential contributors in terms of research output and technological development. Bibliographic coupling and source co-citation analyses further indicated increasing convergence between food science and computational science. Originality\u002Fvalue This study provides a comprehensive overview of the intellectual structure, collaborative networks and thematic evolution of AI-driven food safety research. Unlike previous studies focusing on specific technologies or applications, it maps global research patterns and interdisciplinary linkages. The findings provide strategic insights for researchers, policymakers and industry stakeholders. Future studies should prioritize explainable AI integration into regulatory frameworks and the development of larger, more diverse datasets to improve model generalizability across food systems.","目的 本研究旨在通过全面的文献计量分析，识别人工智能（AI）与食品安全及保障融合领域的最新研究趋势。设计与方法 从Web of Science核心合集中检索了2021年至2025年间发表的2，782篇文献。通过去除重复记录、筛选文献类型并排除不相关记录对数据集进行精炼。随后使用VOSviewer进行合著、关键词共现、引文、共被引及文献耦合分析。研究发现 机器学习和深度学习成为与食品安全、污染检测和质量预测相关的核心技术。关键词聚类显示了与农业、气候变化和公共卫生的跨学科联系。引文和共被引分析识别出三大主要知识基础：传统机器学习、基于计算机视觉的深度学习和时间序列建模。中国、印度和美国在研究产出和技术发展方面是最具影响力的贡献者。文献耦合和来源共被引分析进一步表明食品科学与计算科学之间的融合日益加深。原创性\u002F价值 本研究全面概述了AI驱动的食品安全研究的知识结构、合作网络和主题演变。与以往聚焦特定技术或应用的研究不同，本研究绘制了全球研究模式和跨学科联系。研究结果为研究人员、政策制定者和行业利益相关者提供了战略性见解。未来研究应优先关注可解释AI融入监管框架以及开发更大、更多样化的数据集，以提高模型在食品系统中的泛化能力。","British Food Journal","2026-10-19T00:00:00Z",80,{"impact":19,"substance":215,"depth":19,"authority":20,"freshness":216,"relevant":22,"comment":217},22,8,"基于2782篇文献的AI食品安全计量分析，方法规范、结论有参考价值，但属综述性研究，产业落地影响有限。",[219],{"name":211,"url":208},[27,28,137,30,221],"文献计量",{"code":39,"slug":122,"name":123,"path":124,"parent_code":119},[224],{"code":50,"slug":225,"name":226,"path":227,"parent_code":228},"food-safety","质量安全与溯源","\u002Ftopic\u002Ffood-supply\u002Ffood-safety","E",{"status":126,"locked":14,"version":34,"method":35,"evidence":230},{"candidates":231,"agricultural_evidence":41},[232,237,242,248,255,261],{"code":39,"score":233,"supported":41,"title":234,"abstract":235,"tags":236},19,[139],[46,137,47,138,43,139,44],[28,137],{"code":50,"score":238,"supported":41,"title":239,"abstract":240,"tags":241},15,[144],[30,144],[30],{"code":243,"score":238,"supported":41,"title":244,"abstract":246,"tags":247},"I7",[245],"bibliometric",[221,245],[221],{"code":249,"score":94,"supported":41,"title":250,"abstract":251,"tags":254},"H7",[],[252,253],"气候变化","climate change",[],{"code":256,"score":169,"supported":41,"title":257,"abstract":258,"tags":260},"A8",[],[259],"food security",[],{"code":56,"score":22,"supported":14,"title":262,"abstract":263,"tags":264},[],[],[27],[266,267],"AI 食品安全 文献计量","机器学习 食品质量预测","AI食品安全文献计量-5400","10.1108\u002Fbfj-06-2025-0819",{"doi":269,"openalex_id":271,"authors":272,"venue":211,"cited_by_count":83,"oa_url":9,"card":286,"direction":89,"ingested_from":91},"W7222920742",[273,275,277,280,283],{"name":274,"orcid":9},"Harang Lee",{"name":276,"orcid":9},"지은서",{"name":278,"orcid":279},"Yoonbin Kim","https:\u002F\u002Forcid.org\u002F0000-0002-5625-2560",{"name":281,"orcid":282},"Joon-Goo Lee","https:\u002F\u002Forcid.org\u002F0000-0003-3617-5518",{"name":284,"orcid":285},"Woo‐Ju Kim","https:\u002F\u002Forcid.org\u002F0000-0002-6954-1580",{"tldr":287,"method":288,"finding":289,"direction":89,"opportunity":290},"对2021-2025年2782篇文献做文献计量，梳理AI与食品安全交叉研究趋势。","Web of Science文献计量，VOSviewer共被引、关键词共现与耦合","机器学习与深度学习是核心，形成传统ML、计算机视觉、时序建模三大知识基础。","可解释AI嵌入监管框架、构建更大更多样数据集以提升跨食物系统泛化能力。","2026-10-20T23:30:31.567802Z",{"id":293,"title":294,"url":295,"summary":296,"summary_zh":297,"content":9,"source_name":298,"source_url":295,"published_at":299,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":300,"score_detail":301,"sources":303,"tags":305,"primary_topic":9,"secondary_topics":308,"classification":309,"search_phrases":325,"slug":328,"view_count":51,"doi":329,"paper":330,"created_at":345},4448,"Machine Learning for Ripeness Classification of Nendra Bale Banana: A Classical Approach Using SVM, kNN, and Random Forest","https:\u002F\u002Fdoi.org\u002F10.64388\u002Firev10i4-1723730","acceptance, and overall shelf life. Nendra Bale (Nendran), a dual-purpose, starch-rich Indian banana cultivar widely used for both table consumption and cooking, exhibits a ripening pattern that is markedly more texture-driven than colour-driven, distinguishing it from commercially dominant, colour-dominant cultivars. This study introduces an image-based framework for classifying the ripening stage of Nendra Bale bananas using classical machine learning. A first-hand dataset of 80 raw images across four ripening stages (Green, Mid-ripe, All-Yellow, Overripe) was collected under everyday mobile-camera conditions and augmented to 5,391 images to support more robust model training. Three classical classifiers (Support Vector Machine, k-Nearest Neighbours, Random Forest) were evaluated on Histogram of Oriented Gradients (HOG) features under both raw and augmented conditions. On the raw dataset, kNN gave the highest accuracy (76.47%), ahead of SVM and Random Forest (70.59% each); once the dataset was augmented, SVM became the strongest and most consistent model (81.60%), ahead of kNN (78.62%) and Random Forest (74.53%). A probabilistic stage-to-shelf-life mapping is proposed to convert predictions into an expected remaining life in days. The study offers a low-cost, non-destructive, deployable approach for quality assessment and post-harvest decision support for a texture-dominant, dual-purpose Indian banana cultivar.","接受度以及整体货架期。Nendra Bale（Nendran）是一种兼具鲜食与烹饪用途、富含淀粉的印度香蕉品种，其成熟模式明显更多由质地驱动而非颜色驱动，这使其区别于商业上占主导地位、以颜色为主导的品种。本研究提出了一种基于图像的框架，利用经典机器学习对Nendra Bale香蕉的成熟阶段进行分类。研究在日常手机拍摄条件下收集了涵盖四个成熟阶段（青绿、半熟、全黄、过熟）的80张原始图像一手数据集，并扩充至5,391张图像以支持更稳健的模型训练。在原始和扩充条件下，基于方向梯度直方图（HOG）特征对三种经典分类器（支持向量机、k近邻、随机森林）进行了评估。在原始数据集上，kNN取得了最高准确率（76.47%），高于SVM和随机森林（均为70.59%）；数据集扩充后，SVM成为最强且最一致的模型（81.60%），高于kNN（78.62%）和随机森林（74.53%）。研究提出了一种概率性的阶段至货架期映射方法，将预测结果转换为以天为单位的预期剩余寿命。本研究为一种质地主导、兼具双重用途的印度香蕉品种提供了一种低成本、非破坏性、可部署的品质评估与采后决策支持方法。","Iconic Research and Engineering Journals","2026-10-06T00:00:00Z",55,{"impact":216,"substance":19,"depth":238,"authority":94,"freshness":216,"relevant":22,"comment":302},"面向印度特色香蕉品种的经典机器学习成熟度分级研究，方法常规、数据规模有限，但无损检测与采后决策支持思路对农业信息化有参考价值。",[304],{"name":298,"url":295},[27,28,29,306,307],"香蕉","采后分级",[],{"status":33,"locked":14,"version":34,"method":35,"evidence":310},{"candidates":311,"agricultural_evidence":14,"reason":60},[312,317,321],{"code":39,"score":19,"supported":41,"title":313,"abstract":314,"tags":316},[138],[137,138,315],"decision support",[28],{"code":161,"score":22,"supported":14,"title":318,"abstract":319,"tags":320},[],[],[306],{"code":56,"score":22,"supported":14,"title":322,"abstract":323,"tags":324},[],[],[27],[326,327],"Nendra Bale 香蕉 成熟度 分类","香蕉 成熟度 机器学习 SVM","NendraBale香蕉成熟度分类-4448","10.64388\u002Firev10i4-1723730",{"doi":329,"openalex_id":331,"authors":332,"venue":298,"cited_by_count":83,"oa_url":339,"card":340,"direction":89,"ingested_from":91},"W7220687470",[333,336],{"name":334,"orcid":335},"T. N. Manasa","https:\u002F\u002Forcid.org\u002F0009-0004-1163-2182",{"name":337,"orcid":338},"M Pushpalatha","https:\u002F\u002Forcid.org\u002F0000-0003-3637-9836","https:\u002F\u002Fwww.irejournals.com\u002Fformatedpaper\u002F1723730.pdf",{"tldr":341,"method":342,"finding":343,"direction":89,"opportunity":344},"用HOG特征和经典机器学习对Nendra Bale香蕉四个成熟阶段进行图像分类。","80张原始图像增强至5391张，提取HOG特征，比较SVM、kNN、随机森林。","增强后SVM最优达81.60%，并建立成熟阶段到货架期的概率映射。","可探索纹理主导型品种的深度学习与多模态融合，提升成熟度识别和货架期预测精度。","2026-10-08T23:30:52.769987Z",{"id":347,"title":348,"url":349,"summary":350,"summary_zh":351,"content":9,"source_name":352,"source_url":349,"published_at":353,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":354,"score_detail":355,"sources":357,"tags":359,"primary_topic":9,"secondary_topics":362,"classification":363,"search_phrases":379,"slug":382,"view_count":21,"doi":383,"paper":384,"created_at":415},3133,"Dual-representation fusion of near-infrared spectra for simultaneous identification of fresh-cocoon sex and imprinted-dead cocoons during pupal metamorphosis","https:\u002F\u002Fdoi.org\u002F10.1016\u002Fj.compag.2026.112431","Automated identification of fresh-cocoon sex and imprinted-dead cocoons is essential for raw silk quality control and requires an efficient and reliable approach. Although near-infrared (NIR) spectroscopy shows potential, sex-identification performance across pupal metamorphosis remains insufficiently understood, and existing modeling approaches remain underdeveloped. This study developed a dual-representation weighted-fusion network for fresh-cocoon sex and imprinted-dead cocoon identification during pupal metamorphosis. The network comprises two ablation-optimized branches: a 1D-CNN incorporating dilated spatial attention, adaptive coordinate attention, and ConvNeXtV2 blocks for spectral-sequence feature extraction, and a 2D-CNN incorporating shuffle attention network, adaptive coordinate attention, and ResNet blocks for recurrence-plot-based structural feature extraction. The spectral-sequence and recurrence-plot branches capture complementary information that is integrated using an adaptive weighted-fusion strategy, with the optimized branch parameters used as pretrained weights for transfer learning. On the held-out test set, the dual-representation weighted-fusion network achieved the highest overall accuracy of 95.39%, outperforming the optimized 1D-CNN (93.55%) and 2D-CNN (94.59%), with an average per-sample network inference time of 0.921 ms. Sex-identification performance varied with pupal metamorphosis, showing generally lower and more variable accuracy during the early period, improved performance during the middle period, and some decline during the late period. The fusion network maintained high accuracy across the three periods and achieved its highest mean sex-identification accuracy of 97.22% during the middle period (days 5–7). Overall, the proposed dual-representation NIR framework provides a promising approach for non-destructive fresh-cocoon sex identification and imprinted-dead cocoon screening, while highlighting the pupal developmental period as an important consideration for practical sex sorting.","鲜茧性别与印痕死茧的自动化识别对生丝质量控制至关重要，需要一种高效可靠的方法。尽管近红外（NIR）光谱显示出潜力，但蛹期变态过程中的性别识别性能仍未得到充分理解，现有建模方法也尚不成熟。本研究开发了一种双表示加权融合网络，用于蛹期变态过程中鲜茧性别和印痕死茧的识别。该网络包含两个经消融优化的分支：一个一维CNN分支，融合了膨胀空间注意力、自适应坐标注意力和ConvNeXtV2模块，用于光谱序列特征提取；以及一个二维CNN分支，融合了混洗注意力网络、自适应坐标注意力和ResNet模块，用于基于递归图的结构特征提取。光谱序列分支和递归图分支捕获互补信息，通过自适应加权融合策略进行整合，并将优化后的分支参数用作迁移学习的预训练权重。在留出测试集上，双表示加权融合网络达到了95.39%的最高总体准确率，优于优化后的一维CNN（93.55%）和二维CNN（94.59%），平均单样本网络推理时间为0.921 ms。性别识别性能随蛹期变态而变化，早期准确率普遍较低且波动较大，中期性能提升，晚期有所下降。融合网络在三个时期均保持较高准确率，并在中期（第5–7天）达到了97.22%的最高平均性别识别准确率。总体而言，所提出的双表示近红外框架为非破坏性鲜茧性别识别和印痕死茧筛查提供了一种有前景的方法，同时凸显了蛹发育时期作为实际性别分选的重要考虑因素。","Computers and Electronics in Agriculture","2026-09-21T00:00:00Z",82,{"impact":19,"substance":215,"depth":233,"authority":20,"freshness":107,"relevant":22,"comment":356},"该研究提出双表征加权融合近红外光谱网络，实现鲜茧性别与印痕死茧的高精度无损识别，方法新颖、数据可靠，对蚕桑智能化分选具有实用价值。",[358],{"name":352,"url":349},[27,28,29,360,361],"蚕桑产业","近红外光谱",[],{"status":33,"locked":14,"version":34,"method":35,"evidence":364},{"candidates":365,"agricultural_evidence":14,"reason":60},[366,371,375],{"code":367,"score":22,"supported":14,"title":368,"abstract":369,"tags":370},"A6",[],[],[360],{"code":56,"score":22,"supported":14,"title":372,"abstract":373,"tags":374},[],[],[27],{"code":39,"score":22,"supported":14,"title":376,"abstract":377,"tags":378},[],[],[28],[380,381],"近红外光谱 蚕茧 雌雄鉴别","鲜茧 性别识别 无损检测","近红外光谱蚕茧雌雄鉴别-3133","10.1016\u002Fj.compag.2026.112431",{"doi":383,"openalex_id":385,"authors":386,"venue":352,"cited_by_count":83,"oa_url":349,"card":410,"direction":89,"ingested_from":91},"W7213904433",[387,390,392,394,396,398,400,403,405,408],{"name":388,"orcid":389},"Haibo He","https:\u002F\u002Forcid.org\u002F0000-0002-5247-9370",{"name":391,"orcid":9},"Bing Lu",{"name":393,"orcid":9},"Mingqing Zhong",{"name":395,"orcid":9},"Wantong Xie",{"name":397,"orcid":9},"Yong Chen",{"name":399,"orcid":9},"Zongmeng Yang",{"name":401,"orcid":402},"Hua Huang","https:\u002F\u002Forcid.org\u002F0000-0002-2817-2950",{"name":404,"orcid":9},"Xiaoling Tong",{"name":406,"orcid":407},"Tianfu Zhao","https:\u002F\u002Forcid.org\u002F0000-0001-8189-7638",{"name":409,"orcid":9},"Shiping Zhu",{"tldr":411,"method":412,"finding":413,"direction":89,"opportunity":414},"提出双表征加权融合近红外光谱网络，同时识别鲜茧性别与印痕死茧。","1D-CNN与2D-CNN双分支，结合注意力、递归图与迁移学习。","融合网络总体准确率95.39%，中期性别识别最高达97.22%。","可探索蛹期动态建模与轻量化边缘部署，提升实际分选鲁棒性。","2026-09-22T23:30:02.331612Z",{"id":417,"title":418,"url":419,"summary":420,"summary_zh":421,"content":9,"source_name":352,"source_url":419,"published_at":422,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":423,"score_detail":424,"sources":426,"tags":428,"primary_topic":431,"secondary_topics":432,"classification":433,"search_phrases":450,"slug":453,"view_count":51,"doi":454,"paper":455,"created_at":479},3054,"Assessment of mechanical damage in canola using X-ray radiography and deep learning","https:\u002F\u002Fdoi.org\u002F10.1016\u002Fj.compag.2026.112350","Mechanical damage during post-harvest handling and processing can drastically impair the quality, germination potential, and market value of canola ( Brassica napus L.) seeds. Traditional visual inspection methods are subjective and limited to detecting external defects, often missing internal damage. This study presents a non-destructive approach for classifying mechanical damage in canola seeds using X-ray radiography combined with artificial intelligence techniques. Seeds at three moisture contents (6 %, 8 %, and 12 %) were subjected to four levels of impact energy (0, 3, 3.7, and 4.3 mJ), generating 3,029 radiographic images labelled into four levels of varying mechanical damage (no damage, low damage, medium damage, and high damage). Two analytical pipelines were developed. The first involved machine learning with 37 extracted features, including intensity-based metrics, texture descriptors, frequency-domain features, and Fast Fourier Transform (FFT) features. Support vector machine (SVM) and random forest (RF) models were developed and evaluated; the SVM achieved the highest accuracy of 74.01 % after feature selection. The second pipeline implemented deep learning via transfer learning on five pre-trained convolutional neural networks: MobileNetV2, EfficientNet-B0, Xception, ResNet50, and DenseNet121. Among these, MobileNetV2 achieved the highest accuracy of 91.42 %, outperforming all traditional machine learning models. The results demonstrate the potential of integrating radiographic imaging with artificial intelligence methods for rapid, reliable, and non-destructive classification of mechanical damage in canola seeds, paving the way for intelligent quality control systems in the oilseed supply chain.","油菜（Brassica napus L.）种子在采后处理和加工过程中的机械损伤会严重损害其品质、发芽潜力和市场价值。传统的视觉检测方法主观性强，且仅限于检测外部缺陷，往往无法发现内部损伤。本研究提出了一种利用X射线成像结合人工智能技术对油菜种子机械损伤进行无损分类的方法。将三种含水率（6%、8%和12%）的种子分别施加四个水平的冲击能量（0、3、3.7和4.3 mJ），共生成3,029张射线图像，并标注为四个不同程度的机械损伤等级（无损伤、低损伤、中等损伤和高损伤）。研究开发了两条分析流程。第一条采用机器学习方法，提取了37个特征，包括基于强度的指标、纹理描述符、频域特征以及快速傅里叶变换（FFT）特征，开发并评估了支持向量机（SVM）和随机森林（RF）模型，其中SVM在特征选择后达到了74.01%的最高准确率。第二条流程通过迁移学习在五个预训练卷积神经网络上实现深度学习：MobileNetV2、EfficientNet-B0、Xception、ResNet50和DenseNet121。其中，MobileNetV2达到了91.42%的最高准确率，优于所有传统机器学习模型。结果表明，将射线成像与人工智能方法相结合，有望实现对油菜种子机械损伤的快速、可靠且无损的分类，为油料种子供应链中的智能质量控制系统的建立奠定了基础。","2026-09-20T00:00:00Z",81,{"impact":19,"substance":215,"depth":19,"authority":20,"freshness":107,"relevant":22,"comment":425},"将X射线成像与深度学习结合用于油菜籽内部机械损伤无损分级，方法新颖、数据规模可观，对油料供应链智能质控有参考价值。",[427],{"name":352,"url":419},[27,28,29,429,430],"油菜","种子质量",{"code":39,"slug":122,"name":123,"path":124,"parent_code":119},[],{"status":126,"locked":14,"version":34,"method":35,"evidence":434},{"candidates":435,"agricultural_evidence":41},[436,440,446],{"code":39,"score":19,"supported":41,"title":437,"abstract":438,"tags":439},[43],[46,137,138,43,139],[28],{"code":441,"score":442,"supported":41,"title":443,"abstract":444,"tags":445},"A4",4,[],[429],[429],{"code":56,"score":22,"supported":14,"title":447,"abstract":448,"tags":449},[],[],[27],[451,452],"X射线 油菜籽 机械损伤","MobileNetV2 种子 无损检测","X射线油菜籽机械损伤-3054","10.1016\u002Fj.compag.2026.112350",{"doi":454,"openalex_id":456,"authors":457,"venue":352,"cited_by_count":83,"oa_url":419,"card":474,"direction":89,"ingested_from":91},"W7213778300",[458,460,463,465,468,471],{"name":459,"orcid":9},"M. Buragohain",{"name":461,"orcid":462},"L. G. Divyanth","https:\u002F\u002Forcid.org\u002F0000-0002-1121-1642",{"name":464,"orcid":9},"Taranveer Singh",{"name":466,"orcid":467},"Muhammad Mudassir Arif Chaudhry","https:\u002F\u002Forcid.org\u002F0000-0001-5316-4701",{"name":469,"orcid":470},"Jitendra Paliwal","https:\u002F\u002Forcid.org\u002F0000-0002-1665-3626",{"name":472,"orcid":473},"Mohammad Nadimi","https:\u002F\u002Forcid.org\u002F0000-0002-4550-7572",{"tldr":475,"method":476,"finding":477,"direction":89,"opportunity":478},"用X射线成像结合机器学习与深度学习，无损分类油菜籽机械损伤等级。","3029张X射线图像，提取37个特征训练SVM\u002FRF，并用五种预训练CNN迁移学","MobileNetV2准确率达91.42%，显著优于传统机器学习最高74.01%。","可拓展至多作物种子内部损伤实时检测，并融合高光谱或近红外实现便携式在线分选。","2026-09-21T23:30:02.090869Z",{"id":481,"title":482,"url":483,"summary":484,"summary_zh":485,"content":9,"source_name":486,"source_url":483,"published_at":487,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":15,"score_detail":488,"sources":492,"tags":494,"primary_topic":497,"secondary_topics":498,"classification":499,"search_phrases":520,"slug":523,"view_count":51,"doi":524,"paper":525,"created_at":541},3017,"Automated seedling vigor estimation in cucumbers using digital image processing","https:\u002F\u002Fdoi.org\u002F10.65764\u002Ftjas.2026.267412","Background and Objective: Farmers often rely on experience rather than quantitative indicators to determine seedling quality, resulting in inefficiencies in resource allocation. This study aimed to (1) compare seedling vigor and growth characteristics between open-pollinated (OP) and F1 hybrid cucumber seed types, and (2) develop a non-destructive predictive model for seedling vigor estimation based on digital image analysis.Methodology: Seedling images were acquired under controlled LED lighting using a 5 MP digital camera positioned 30 cm above 14-day-old seedlings. A YOLOv8-based object detection model was applied to detect and isolate true leaf regions, from which pixel count and RGB color values were extracted as input features for a multiple linear regression model to predict the seedling vigor index (SVI).Main Results: The YOLOv8 object detection model achieved a mean average precision (mAP@0.5) of 96.00%, precision of 93.40%, and recall of 94.40% in detecting true leaves. Multiple linear regression analysis was conducted using the Scikit-learn library in Python. Scikit-learn provides regression-based machine learning algorithms, including multiple linear regression. The equation was then applied to the training and test sets, using the pixel values of true leaves as the independent variable and SVI as the dependent variable. Using multiple regression analysis, the trained model generated the equation SVI = -2.27 + (2.84 × 10-5 × pixel) + (3.48 × 10-5 × R) + (-7.49 × 10-2 × G) + (2.06 × 10-1 × B), R2 = 0.57 and RMSE = 1.10. The model achieved the test set, r = 0.79 and RMSE = 1.34. The model performance showed a correlation coefficient of 0.85 for OP data and 0.91 for F1 hybrid data.Conclusions: The results demonstrate that digital image processing combined with object detection provides a non-destructive and effective approach to estimate seedling vigor quality. The predictive models developed from OP and F1 hybrid datasets indicate potential application in precision agriculture for automated seedling quality assessment and transplanting decision support.","背景与目标：农民通常依赖经验而非定量指标来判断幼苗质量，导致资源配置效率低下。本研究旨在（1）比较开放授粉（OP）与F1杂交黄瓜种子类型之间的幼苗活力和生长特性，（2）基于数字图像分析开发一种用于幼苗活力评估的无损预测模型。方法：使用500万像素数码相机置于14日龄幼苗上方30 cm处，在受控LED光照下获取幼苗图像。应用基于YOLOv8的目标检测模型检测并分离真叶区域，从中提取像素计数和RGB颜色值作为多元线性回归模型的输入特征，以预测幼苗活力指数（SVI）。主要结果：YOLOv8目标检测模型在检测真叶时达到了96.00%的平均精度均值（mAP@0.5）、93.40%的精确率和94.40%的召回率。使用Python中的Scikit-learn库进行多元线性回归分析。Scikit-learn提供基于回归的机器学习算法，包括多元线性回归。随后将该方程应用于训练集和测试集，以真叶像素值作为自变量，SVI作为因变量。通过多元回归分析，训练模型生成的方程为SVI = -2.27 + (2.84 × 10-5 × 像素) + (3.48 × 10-5 × R) + (-7.49 × 10-2 × G) + (2.06 × 10-1 × B)，R2 = 0.57，RMSE = 1.10。该模型在测试集上达到r = 0.79，RMSE = 1.34。模型性能显示，OP数据的相关系数为0.85，F1杂交数据的相关系数为0.91。结论：结果表明，数字图像处理结合目标检测为评估幼苗活力质量提供了一种无损且有效的方法。基于OP和F1杂交数据集开发的预测模型表明，其在精准农业中具有用于自动化幼苗质量评估和移栽决策支持的潜在应用。","Thai Journal of Agricultural Science","2026-09-19T00:00:00Z",{"impact":17,"substance":489,"depth":490,"authority":17,"freshness":107,"relevant":22,"comment":491},21,17,"基于YOLOv8与多元回归的黄瓜幼苗活力无损估测，方法具体、指标完整，对智慧育苗有参考价值，但属细分作物研究，影响范围有限。",[493],{"name":486,"url":483},[27,28,29,495,496],"图像识别","黄瓜育苗",{"code":39,"slug":122,"name":123,"path":124,"parent_code":119},[],{"status":126,"locked":14,"version":34,"method":35,"evidence":500},{"candidates":501,"agricultural_evidence":41},[502,508,514],{"code":39,"score":503,"supported":41,"title":504,"abstract":505,"tags":507},11,[],[506,137,138,315],"目标检测",[28,495],{"code":56,"score":21,"supported":41,"title":509,"abstract":510,"tags":513},[],[511,512],"精准农业","precision agriculture",[27],{"code":154,"score":94,"supported":41,"title":515,"abstract":516,"tags":519},[],[517,518],"黄瓜","cucumber",[],[521,522],"黄瓜 幼苗活力 图像处理","YOLOv8 幼苗 检测","黄瓜幼苗活力图像处理-3017","10.65764\u002Ftjas.2026.267412",{"doi":524,"openalex_id":526,"authors":527,"venue":486,"cited_by_count":83,"oa_url":483,"card":536,"direction":203,"ingested_from":91},"W7213631983",[528,530,532,534],{"name":529,"orcid":9},"Thanabodee Withunchettanan",{"name":531,"orcid":9},"Raksak Sermsak",{"name":533,"orcid":9},"Pichittra Kaewsorn",{"name":535,"orcid":9},"Kriengkri Kaewtrakulpong",{"tldr":537,"method":538,"finding":539,"direction":89,"opportunity":540},"用YOLOv8检测黄瓜真叶并结合多元回归，实现幼苗活力指数无损预测。","LED下拍摄14天幼苗，YOLOv8分割真叶，提取像素与RGB做多元线性回归。","YOLOv8检测mAP@0.5达96%，模型测试r=0.79，F1杂交种相关性达0.91。","可扩展多品种、多环境数据，融合时序图像与深度学习提升活力预测泛化性。","2026-09-20T23:30:26.929607Z"]