[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-4580":3,"related-4580":66},{"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":18,"tags":20,"primary_topic":9,"secondary_topics":24,"classification":25,"search_phrases":40,"slug":43,"view_count":15,"doi":44,"paper":45,"created_at":65},4580,"Phage phenotyping by measuring plaque expansion dynamics","https:\u002F\u002Fdoi.org\u002F10.64898\u002F2026.10.07.757331","Bacteriophages exhibit staggering genomic diversity, yet phenotypic characterization remains bottlenecked by assays that generally yield simple binary infection outcomes. Quantitative traits can be extracted from standard liquid cultures, but many phages fail to produce a measurable effect in these settings, necessitating an improved quantitative approach on semi-solid media. While end-point plaque size is commonly used for quantitative phenotyping, it only provides a static snapshot that fails to disentangle phage performance across changing bacterial growth phases. Here, we report a cost-effective, scalable method to quantify the bacterial killing rates of diverse phages by continuously tracking plaque development using a consumer-grade flatbed scanner and custom open-source computer vision software. By fitting a minimal, five-parameter phenomenological model to the diverse BASEL collection of Escherichia coli phages, we accurately captured plaque expansion dynamics across all bacterial growth phases. We discovered that plaque expansion parameters vary significantly across individual phages and taxonomic families, with the initial expansion rate demonstrating an inverse correlation with viral genome size. Notably, these plaque expansion dynamics did not correlate with bacterial collapse times in liquid culture, indicating that they represent distinct physiological and kinetic processes. Finally, we demonstrated that phage taxonomic family can be predicted with 72.1% accuracy solely from dynamic plaque parameters using a random forest classifier, whereas conventional endpoint plaque sizes held no predictive power. This time-lapse phenotyping approach yields robust kinetic data superior to end-point measurements, establishing a scalable platform to accelerate phage discovery and genotype-phenotype mapping.","噬菌体展现出惊人的基因组多样性，然而其表型特征鉴定仍受限于通常只能给出简单二元感染结果的测定方法。定量性状可以从标准液体培养中提取，但许多噬菌体在这些条件下无法产生可测量的效应，因此需要在半固体培养基上建立改进的定量方法。虽然终点噬菌斑大小常用于定量表型分析，但它仅提供静态快照，无法区分噬菌体在细菌不同生长阶段中的表现。在此，我们报道了一种经济高效、可扩展的方法，通过使用消费级平板扫描仪和定制的开源计算机视觉软件连续追踪噬菌斑发育，来量化多种噬菌体的细菌杀灭速率。通过将简约的五参数现象学模型拟合到多样化的BASEL大肠杆菌（Escherichia coli）噬菌体集合，我们准确捕捉了所有细菌生长阶段的噬菌斑扩展动态。我们发现，噬菌斑扩展参数在单个噬菌体和分类科之间存在显著差异，其中初始扩展速率与病毒基因组大小呈负相关。值得注意的是，这些噬菌斑扩展动态与液体培养中的细菌崩溃时间不相关，表明它们代表了不同的生理和动力学过程。最后，我们证明，仅使用动态噬菌斑参数，通过随机森林分类器即可预测噬菌体分类科，准确率达72.1%，而传统终点噬菌斑大小则不具备预测能力。这种延时表型分析方法产生的稳健动力学数据优于终点测量，建立了一个可扩展的平台，以加速噬菌体发现和基因型-表型映射。",null,"bioRxiv (Cold Spring Harbor Laboratory)","2026-10-08T00:00:00Z","论文",10,false,0,{"impact":15,"substance":15,"depth":15,"authority":15,"freshness":15,"relevant":15,"comment":17},"噬菌体表型定量方法属基础微生物学，与三农、农业信息化无直接关联，不建议进入每日精选。",[19],{"name":10,"url":6},[21,22,23],"计算机视觉","噬菌体表型","微生物检测",[],{"status":26,"locked":14,"version":27,"method":28,"evidence":29},"needs_review","2026-10-22.2","rules",{"candidates":30,"agricultural_evidence":14,"reason":39},[31],{"code":32,"score":33,"supported":34,"title":35,"abstract":36,"tags":38},"I2",7,true,[],[21,37],"computer vision",[21],"农业相关性需要复核",[41,42],"噬菌体 表型 平板 扫描仪","噬菌体表型 微生物检测 计算机视觉","噬菌体表型平板扫描仪-4580","10.64898\u002F2026.10.07.757331",{"doi":44,"openalex_id":46,"authors":47,"venue":10,"cited_by_count":15,"oa_url":56,"card":57,"direction":63,"ingested_from":64},"W7221007543",[48,50,53],{"name":49,"orcid":9},"Jonathan Costa",{"name":51,"orcid":52},"Baltus A. van der Steen","https:\u002F\u002Forcid.org\u002F0000-0001-8193-4814",{"name":54,"orcid":55},"Yuval Mulla","https:\u002F\u002Forcid.org\u002F0000-0002-6675-0977","https:\u002F\u002Fwww.biorxiv.org\u002Fcontent\u002Fbiorxiv\u002Fearly\u002F2026\u002F10\u002F08\u002F2026.10.07.757331.full.pdf",{"tldr":58,"method":59,"finding":60,"direction":61,"opportunity":62},"用平板扫描仪连续追踪噬菌体噬菌斑扩展，建立定量表型新方法。","消费级扫描仪+开源计算机视觉，拟合五参数模型分析大肠杆菌噬菌体。","噬菌斑扩展参数因噬菌体而异，初始扩展率与基因组大小负相关，可预测分类科。","其他","将时间序列表型分析引入农业噬菌体防治，挖掘噬菌体-宿主互作动态与田间应用潜力。","农业遥感与作物表型","openalex","2026-10-10T23:30:25.274181Z",{"total":67,"page":68,"page_size":67,"items":69},6,1,[70,139,193,237,295,336],{"id":71,"title":72,"url":73,"summary":74,"summary_zh":75,"content":9,"source_name":76,"source_url":73,"published_at":77,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":78,"score_detail":79,"sources":85,"tags":87,"primary_topic":92,"secondary_topics":97,"classification":98,"search_phrases":118,"slug":121,"view_count":15,"doi":122,"paper":123,"created_at":138},5587,"Calibration-Shift Inference for Computer Vision Agriculture Yield Models Under Retrofit Sensor Replacement: An Artificial Intelligence Post-Deployment Recalibration Guarantee","https:\u002F\u002Fdoi.org\u002F10.5281\u002Fzenodo.23544195","Computer vision yield models deployed in agricultural fields are routinely paired with retrofit sensor packages that replace or supplement the original imaging hardware, and the resulting calibration shift can silently invalidate the model's predictive guarantees. The analysis characterises the operating characteristics of the design under a declared variance structure with between-cluster standard deviation 6.0, residual standard deviation 4.5 and intraclass correlation 0.64, using 2000 replications per design cell across cluster counts of 8, 16, 24 and 32 and effect sizes of 3, 5 and 8 percentage points. The design attains 80% power at a 5 percentage point effect with 54 clusters. The paper states a formal guarantee: under stated assumptions on the stability of the calibration map, the before-after-control-impact estimator yields a consistent and asymptotically normal estimate of the calibration shift, and the associated confidence interval covers the true shift at the nominal rate. The contribution is a design-theoretic account of how retrofit-induced calibration shift can be detected and bounded without observational data, together with the sample-size requirements that follow.","部署在农田中的计算机视觉产量模型通常与改装传感器组件配合使用，这些组件用于替换或补充原始成像硬件，由此产生的校准偏移可能悄然使模型的预测保证失效。该分析在声明的方差结构下刻画了设计的工作特性，其中组间标准差为6.0、残差标准差为4.5、组内相关系数为0.64，每个设计单元采用2000次重复，聚类数分别为8、16、24和32，效应量为3、5和8个百分点。该设计在54个聚类、5个百分点效应下达到80%的检验力。论文给出了一项形式化保证：在关于校准映射稳定性的既定假设下，前后对照影响（before-after-control-impact）估计量给出校准偏移的一致且渐近正态的估计，且相关置信区间以名义覆盖率覆盖真实偏移。其贡献在于从设计理论角度说明了如何在没有观测数据的情况下检测并界定改装引起的校准偏移，以及随之而来的样本量要求。","Zenodo (CERN European Organization for Nuclear Research)","2026-10-21T00:00:00Z",65,{"impact":80,"substance":81,"depth":82,"authority":83,"freshness":13,"relevant":68,"comment":84},8,18,17,12,"方法学扎实、统计设计严谨的预印本论文，但偏理论设计层面，缺乏田间实证数据，产业影响有限，适合作为技术参考而非每日精选头条。",[86],{"name":76,"url":73},[88,89,21,90,91],"智慧农业","农业人工智能","产量预测","传感器校准",{"code":32,"slug":93,"name":94,"path":95,"parent_code":96},"agricultural-ai","农业AI与大模型","\u002Ftopic\u002Fdigital-agriculture\u002Fagricultural-ai","I",[],{"status":99,"locked":14,"version":27,"method":28,"evidence":100},"classified",{"candidates":101,"agricultural_evidence":34},[102,108,113],{"code":32,"score":103,"supported":34,"title":104,"abstract":106,"tags":107},24,[105,37],"artificial intelligence",[21,37],[89,21],{"code":109,"score":68,"supported":14,"title":110,"abstract":111,"tags":112},"I1",[],[],[88],{"code":114,"score":68,"supported":14,"title":115,"abstract":116,"tags":117},"I3",[],[],[90],[119,120],"计算机视觉 产量预测 传感器校准","农业人工智能 传感器校准 计算机视觉 产量预测","计算机视觉产量预测传感器校准-5587","10.5281\u002Fzenodo.23544195",{"doi":122,"openalex_id":124,"authors":125,"venue":76,"cited_by_count":15,"oa_url":73,"card":132,"direction":136,"ingested_from":64},"W7223104380",[126,128,130],{"name":127,"orcid":9},"Tigist Alemu Gebremariam",{"name":129,"orcid":9},"Hawi Tesfaye Desta",{"name":131,"orcid":9},"Mulugeta Bekele Wolde",{"tldr":133,"method":134,"finding":135,"direction":136,"opportunity":137},"提出改造传感器替换后视觉产量模型校准漂移的检测与置信区间保证方法。","用前后对照影响估计量，基于方差结构做2000次重复的模拟设计分析。","5个百分点效应下54个集群可达80%检验力，估计量一致且渐近正态。","农业人工智能与决策模型","可将该设计理论扩展到多传感器异构替换与在线校准监测，填补真实田间验证空白。","2026-10-22T11:21:01.554995Z",{"id":140,"title":141,"url":142,"summary":143,"summary_zh":144,"content":9,"source_name":76,"source_url":142,"published_at":77,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":145,"score_detail":146,"sources":150,"tags":152,"primary_topic":155,"secondary_topics":156,"classification":157,"search_phrases":173,"slug":176,"view_count":15,"doi":177,"paper":178,"created_at":192},5584,"Distributionally-Robust Evaluation of Computer Vision Agriculture Disease Classifiers Under Climate-Driven Pathogen Range Shifts: An Artificial Intelligence Worst-Case Generalization Bound","https:\u002F\u002Fdoi.org\u002F10.5281\u002Fzenodo.23544150","Computer vision classifiers for crop disease are evaluated almost exclusively on held-out samples drawn from the same distribution as their training data, yet climate-driven pathogen range shifts mean that deployment environments are systematically out of distribution. The framework treats the deployment distribution as an adversarial perturbation of the training distribution within a Wasserstein ball whose radius is calibrated to observed climate displacement. Using NASA POWER monthly reanalysis for Harare, Zimbabwe (2000-2023, 312 monthly observations), the analysis estimates a mean air temperature of 18.914 C, a standard deviation of 3.107 C, and a warming trend of 0.2001 C per decade. These estimates anchor the radius of the ambiguity set and the resulting bound. The bound decomposes worst-case risk into an empirical risk term, a Wasserstein radius term scaled by the Lipschitz constant of the classifier, and a concentration term. Under plausible Lipschitz constants for convolutional backbones, the radius term dominates the empirical risk term once the climate displacement exceeds a threshold that the Harare record already approaches. The contribution is a formal evaluation criterion that replaces the i. i. d. assumption with a climate-calibrated ambiguity set, together with the analytical machinery needed to compute it.","作物病害的计算机视觉分类器几乎完全在与训练数据同分布的留出样本上进行评估，然而气候驱动的病原体分布范围变化意味着部署环境系统性地处于分布之外。该框架将部署分布视为训练分布在Wasserstein球内的对抗性扰动，球的半径根据观测到的气候位移进行校准。利用NASA POWER对津巴布韦哈拉雷的月度再分析数据（2000—2023年，312个月度观测值），分析估计平均气温为18.914°C，标准差为3.107°C，每十年升温趋势为0.2001°C。这些估计值锚定了模糊集的半径及由此得到的界。该界将最坏情况风险分解为经验风险项、由分类器Lipschitz常数缩放后的Wasserstein半径项以及集中项。在卷积骨干网络合理的Lipschitz常数下，一旦气候位移超过哈拉雷记录已接近的阈值，半径项便主导经验风险项。本文的贡献在于提出了一种形式化评估准则，以气候校准的模糊集替代独立同分布（i.i.d.）假设，并提供了计算该准则所需的分析工具。",75,{"impact":83,"substance":147,"depth":81,"authority":148,"freshness":13,"relevant":68,"comment":149},22,13,"方法新颖、数据具体，但属方法论预印论文，产业影响有限，适合作为专业精选而非头条。",[151],{"name":76,"url":142},[88,89,21,153,154],"作物病害","气候变化",{"code":32,"slug":93,"name":94,"path":95,"parent_code":96},[],{"status":99,"locked":14,"version":27,"method":28,"evidence":158},{"candidates":159,"agricultural_evidence":34},[160,164,169],{"code":32,"score":103,"supported":34,"title":161,"abstract":162,"tags":163},[105,37],[21,37],[89,21],{"code":165,"score":68,"supported":14,"title":166,"abstract":167,"tags":168},"H7",[],[],[154],{"code":109,"score":68,"supported":14,"title":170,"abstract":171,"tags":172},[],[],[88],[174,175],"NASA POWER 哈拉雷 气温","Wasserstein 分布鲁棒 病害分类","NASAPOWER哈拉雷气温-5584","10.5281\u002Fzenodo.23544150",{"doi":177,"openalex_id":179,"authors":180,"venue":76,"cited_by_count":15,"oa_url":142,"card":187,"direction":136,"ingested_from":64},"W7223087892",[181,183,185],{"name":182,"orcid":9},"Kossi Afeli",{"name":184,"orcid":9},"Akouvi Amegashie",{"name":186,"orcid":9},"Komlan Sélom Dossou",{"tldr":188,"method":189,"finding":190,"direction":136,"opportunity":191},"提出气候校准的分布鲁棒评估准则，替代i.i.d.假设评价作物病害视觉分类器。","用Wasserstein球建模部署分布偏移，以NASA POWER哈拉雷气温数据","气候位移超阈值后，半径项主导经验风险，现有评估会高估模型泛化能力。","可将气候校准的鲁棒评估扩展到多作物多区域，并研究适应分布偏移的鲁棒训练方法。","2026-10-22T11:21:00.036937Z",{"id":194,"title":195,"url":196,"summary":197,"summary_zh":198,"content":9,"source_name":199,"source_url":196,"published_at":200,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":15,"score_detail":201,"sources":203,"tags":205,"primary_topic":9,"secondary_topics":207,"classification":208,"search_phrases":218,"slug":221,"view_count":15,"doi":222,"paper":223,"created_at":236},5494,"YOLO-Based Real-Time Object Detection Using Computer Vision for Robotics","https:\u002F\u002Fdoi.org\u002F10.64388\u002Firev10i4-1724117","Robots operating in dynamic environments require rapid and reliable perception to identify people, tools, obstacles, and other objects before making movement or manipulation decisions. Conventional vision pipelines may require separate stages for region proposal, classification, and localization, which can increase computational cost and latency. You Only Look Once (YOLO) formulates object detection as a single-network prediction task and has become a practical family of detectors for real-time computer-vision applications. This paper presents a research-oriented framework for integrating a lightweight YOLO detector with a robot-mounted camera. The framework covers image acquisition, preprocessing, object detection, confidence filtering, coordinate interpretation, and communication with a robot controller. It also discusses model selection, deployment constraints, evaluation metrics, safety considerations, and a reproducible testing plan. Rather than reporting unperformed experiments, the paper defines how a prototype should be evaluated using precision, recall, mean average precision, inference latency, frame rate, and end-to-end response time. The analysis indicates that compact YOLO variants are suitable starting points for resource-constrained robotic platforms, while reliable navigation requires additional depth or distance sensing and a safety controller. The proposed approach can support obstacle awareness, object following, pick-and-place assistance, and indoor mobile-robot perception.","在动态环境中运行的机器人需要快速且可靠的感知能力，以便在做出移动或操作决策之前识别人、工具、障碍物和其他物体。传统视觉流水线可能需要区域提议、分类和定位等独立阶段，这会增加计算成本和延迟。You Only Look Once（YOLO）将目标检测表述为单网络预测任务，已成为实时计算机视觉应用中实用的一类检测器。本文提出了一个面向研究的框架，用于将轻量级YOLO检测器与机器人搭载的相机集成。该框架涵盖图像采集、预处理、目标检测、置信度过滤、坐标解释以及与机器人控制器的通信。文中还讨论了模型选择、部署约束、评估指标、安全考虑以及可复现的测试方案。本文不报告未进行的实验，而是定义了应如何使用精确率、召回率、平均精度均值、推理延迟、帧率和端到端响应时间来评估原型。分析表明，紧凑型YOLO变体适合作为资源受限机器人平台的起点，而可靠导航还需要额外的深度或距离传感以及安全控制器。所提出的方法可支持障碍物感知、目标跟随、拾取与放置辅助以及室内移动机器人感知。","Iconic Research and Engineering Journals","2026-10-19T00:00:00Z",{"impact":15,"substance":15,"depth":15,"authority":15,"freshness":15,"relevant":15,"comment":202},"该文为通用机器人视觉论文，未涉及农业场景，与三农及农业信息化无直接关联，不建议进入每日精选。",[204],{"name":199,"url":196},[89,21,206],"机器人感知",[],{"status":26,"locked":14,"version":27,"method":28,"evidence":209},{"candidates":210,"agricultural_evidence":14,"reason":39},[211],{"code":32,"score":212,"supported":34,"title":213,"abstract":215,"tags":217},27,[37,214],"YOLO",[216,21,214],"目标检测",[89,21],[219,220],"YOLO 机器人 目标检测","农业人工智能 机器人感知 计算机视觉","YOLO机器人目标检测-5494","10.64388\u002Firev10i4-1724117",{"doi":222,"openalex_id":224,"authors":225,"venue":199,"cited_by_count":15,"oa_url":230,"card":231,"direction":136,"ingested_from":64},"W7222933754",[226,228],{"name":227,"orcid":9},"Aafareen Sikandar Shaikh",{"name":229,"orcid":9},"Archana Nilesh Pawar","https:\u002F\u002Fwww.irejournals.com\u002Fformatedpaper\u002F1724117.pdf",{"tldr":232,"method":233,"finding":234,"direction":136,"opportunity":235},"提出将轻量YOLO检测器与机器人相机集成的框架，并定义原型评估方案。","轻量YOLO单阶段检测、机器人相机图像采集与预处理、置信度过滤及控制器通信。","紧凑YOLO适合资源受限机器人平台，但可靠导航仍需深度感知与安全控制器。","可将该框架迁移至农业机器人，研究田间遮挡与光照变化下的轻量检测与深度感知融合。","2026-10-21T23:30:40.721716Z",{"id":238,"title":239,"url":240,"summary":241,"summary_zh":242,"content":9,"source_name":243,"source_url":240,"published_at":244,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":245,"score_detail":246,"sources":249,"tags":251,"primary_topic":254,"secondary_topics":255,"classification":256,"search_phrases":280,"slug":283,"view_count":15,"doi":284,"paper":285,"created_at":294},5385,"A Comprehensive Review of Deep Learning and Computer Vision Techniques for Fruit Quality Assessment and Classification","https:\u002F\u002Fdoi.org\u002F10.25258\u002Fijddt.16.75s.151","Fruit image analysis has become an essential research area in precision agriculture, enabling automated fruit classification, quality grading, maturity estimation, disease diagnosis, yield prediction, and intelligent harvesting through advances in computer vision and artificial intelligence (AI).Recent progress in imaging technologies, large-scale benchmark datasets, and deep learning has substantially improved the accuracy, robustness, and scalability of automated fruit analysis systems.This survey presents a comprehensive and systematic review of image processing and AI techniques for fruit image analysis.It covers the complete image analysis pipeline, including image acquisition, preprocessing, enhancement, color space transformation, segmentation, feature extraction, and handcrafted texture descriptors.Furthermore, conventional machine learning algorithms are reviewed alongside state-of-the-art deep learning architectures, including Convolutional Neural Networks (CNNs), Vision Transformers (ViTs), and Vision Mamba models, for applications such as fruit recognition, quality assessment, disease detection, ripeness estimation, and object detection.Publicly available benchmark datasets, evaluation metrics, and performance comparison strategies are also summarized to facilitate objective assessment of existing approaches.In addition, the survey discusses key research challenges, including illumination variations, occlusion, complex orchard environments, limited annotated datasets, computational complexity, and domain adaptation, while highlighting emerging research trends such as self-supervised learning, multimodal learning, explainable AI, lightweight deep networks, foundation models, and edge intelligence for real-time deployment.By providing a unified review of methodologies, datasets, evaluation protocols, challenges, and future research directions, this survey offers a comprehensive reference for researchers and practitioners developing intelligent and scalable fruit image analysis systems for precision agriculture.","水果图像分析已成为精准农业中的一个重要研究领域，借助计算机视觉和人工智能（AI）的进步，实现了果实自动分类、品质分级、成熟度估计、病害诊断、产量预测和智能采收。成像技术、大规模基准数据集和深度学习的最新进展显著提升了自动化水果分析系统的准确性、鲁棒性和可扩展性。本综述对用于水果图像分析的图像处理和AI技术进行了全面而系统的回顾。内容涵盖完整的图像分析流程，包括图像采集、预处理、增强、颜色空间变换、分割、特征提取和手工纹理描述符。此外，综述了传统机器学习算法以及最先进的深度学习架构，包括卷积神经网络（CNN）、视觉Transformer（ViT）和视觉Mamba模型，用于果实识别、品质评估、病害检测、成熟度估计和目标检测等应用。同时总结了公开可用的基准数据集、评价指标和性能比较策略，以便对现有方法进行客观评估。此外，本综述讨论了关键研究挑战，包括光照变化、遮挡、复杂果园环境、标注数据集有限、计算复杂度和域适应，并重点介绍了新兴研究趋势，如自监督学习、多模态学习、可解释AI、轻量级深度网络、基础模型和面向实时部署的边缘智能。通过提供方法、数据集、评价协议、挑战和未来研究方向的统一综述，本综述为开发面向精准农业的智能且可扩展的水果图像分析系统的研究人员和实践者提供了全面的参考。","International Journal of Drug Delivery Technology","2026-10-18T00:00:00Z",56,{"impact":80,"substance":81,"depth":247,"authority":67,"freshness":80,"relevant":68,"comment":248},16,"系统性综述覆盖CNN、ViT、Vision Mamba等前沿方法，信息量足，但期刊权威性一般且偏学术综述，公共价值有限，适合主题聚合而非每日精选头条。",[250],{"name":243,"url":240},[88,89,21,252,253],"深度学习","水果品质检测",{"code":32,"slug":93,"name":94,"path":95,"parent_code":96},[],{"status":99,"locked":14,"version":27,"method":28,"evidence":257},{"candidates":258,"agricultural_evidence":34},[259,267,273],{"code":32,"score":260,"supported":34,"title":261,"abstract":263,"tags":266},28,[262,37],"deep learning",[216,252,264,21,265,262,105,37],"机器学习","machine learning",[89,252,21],{"code":109,"score":33,"supported":34,"title":268,"abstract":269,"tags":272},[],[270,271],"精准农业","precision agriculture",[88],{"code":274,"score":275,"supported":34,"title":276,"abstract":277,"tags":279},"B2",3,[],[278],"orchard",[],[281,282],"水果品质 计算机视觉 深度学习","农业人工智能 水果品质检测 计算机视觉 智慧农业","水果品质计算机视觉深度学习-5385","10.25258\u002Fijddt.16.75s.151",{"doi":284,"openalex_id":286,"authors":287,"venue":243,"cited_by_count":15,"oa_url":240,"card":288,"direction":293,"ingested_from":64},"W7222892831",[],{"tldr":289,"method":290,"finding":291,"direction":63,"opportunity":292},"系统综述深度学习与计算机视觉在水果品质评估与分级中的全流程方法。","综述图像采集到分割、特征提取及CNN、ViT、Vision Mamba等模型与公","深度学习显著提升水果分析精度与鲁棒性，但光照、遮挡和小样本仍是主要瓶颈。","可探索自监督与轻量化模型在复杂果园遮挡、小样本场景下的边缘实时部署。","智慧农业 \u002F 农业物联网","2026-10-20T23:30:10.066042Z",{"id":296,"title":297,"url":298,"summary":299,"summary_zh":300,"content":9,"source_name":301,"source_url":298,"published_at":244,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":15,"score_detail":302,"sources":304,"tags":306,"primary_topic":9,"secondary_topics":309,"classification":310,"search_phrases":317,"slug":320,"view_count":15,"doi":321,"paper":322,"created_at":335},5384,"Food Recognition and Nutrition Estimator Using Computer Vision and Machine Learning","https:\u002F\u002Fdoi.org\u002F10.32628\u002Fcseit2612569","Food logging works when it is simple enough to become a habit. Traditional nutrition programs often require the user to browse a list of foods, choose a food, and guess at a portion size, and then repeat the procedure for each food item in a meal. CrunchUP resolves that friction by using a photo of a meal to create a nutrition record. The system is a complete web application, where a user uploads an image or takes a picture using their camera, a FastAPI service checks and standardizes the image, and the Google Gemini vision model identifies the foods visible in the image and estimates the quantity, calories, protein, carbohydrates, fat, and fibre. The response is validated, converted into application objects, displayed to the user and optionally stored in PostgreSQL. The information on a personal profile includes height, weight, age, gender, activity level, and fitness goal. These values are used for the calculation of BMI, basal metabolic rate according to Mifflin–St Jeor, total daily energy expenditure, calorie adjustment which is goal oriented and macronutrient targets. The work is important because it brings together recognition, nutrition estimation, personal targets, meal history, trends and feedback for decision making in a single stream. The report describes the requirements, literature background, architecture, database design, algorithms, security measures, implementation, verification strategy, restrictions and extensions for the future. Separation of concerns is evident in the repository: the client is responsive (React, Vite); the service and data are implemented (FastAPI, Pydantic, SQLAlchemy); access is protected (JWT, bcrypt); images are normalized (Pillow); and multimodal output is schema shaped (Gemini 2.5 Flash). The dashboard consolidates intake and translates progress data into easy-to-understand recommendations. A key finding is that a well-defined vision-language service can help ease recording effort in the diet, but needs to be interpreted as guidance, not a clinical measurement. There is a need for future work including calibrated portion references, verified food databases, confidence reporting, culturally wider evaluation, personalised meal recommendation, and the development of a mobile app in the native application platform \"React Native\".","食物记录只有在足够简单、能够成为习惯时才会奏效。传统的营养项目往往要求用户浏览食物列表、选择一种食物、猜测份量，然后对一餐中的每一种食物重复这一过程。CrunchUP通过使用一张餐食照片来创建营养记录，从而解决了这一摩擦。该系统是一个完整的Web应用程序：用户上传图片或使用相机拍照，FastAPI服务对图像进行检查和标准化，Google Gemini视觉模型识别图像中可见的食物，并估算数量、热量、蛋白质、碳水化合物、脂肪和纤维。响应经过验证，转换为应用程序对象，展示给用户，并可选地存储于PostgreSQL中。个人档案中的信息包括身高、体重、年龄、性别、活动水平和健身目标。这些数值用于计算BMI、依据Mifflin–St Jeor公式计算基础代谢率、每日总能量消耗、以目标为导向的热量调整以及宏量营养素目标。这项工作之所以重要，是因为它将识别、营养估算、个人目标、餐食历史、趋势和用于决策的反馈整合到单一流程中。本报告描述了需求、文献背景、架构、数据库设计、算法、安全措施、实现、验证策略、限制以及未来的扩展。关注点分离在代码库中清晰可见：客户端具有响应式设计（React、Vite）；服务与数据层已实现（FastAPI、Pydantic、SQLAlchemy）；访问受到保护（JWT、bcrypt）；图像经过标准化处理（Pillow）；多模态输出按模式塑形（Gemini 2.5 Flash）。仪表盘整合摄入信息，并将进度数据转化为易于理解的建议。一个关键发现是，定义良好的视觉-语言服务有助于减轻饮食记录负担，但必须将其解读为参考指导，而非临床测量。未来工作需要包括校准份量参照、经验证的食物数据库、置信度报告、文化上更广泛的评估、个性化餐食推荐，以及开发基于原生应用平台“React Native”的移动应用。","International Journal of Scientific Research in Computer Science Engineering and Information Technology",{"impact":15,"substance":15,"depth":15,"authority":15,"freshness":15,"relevant":15,"comment":303},"该文介绍的是面向个人膳食记录的食物识别与营养估算应用，属于通用健康信息化，与三农、农业信息化、智慧农业无直接关联，不建议进入每日精选。",[305],{"name":301,"url":298},[89,21,307,308],"营养估算","膳食记录",[],{"status":26,"locked":14,"version":27,"method":28,"evidence":311},{"candidates":312,"agricultural_evidence":14,"reason":39},[313],{"code":32,"score":81,"supported":34,"title":314,"abstract":315,"tags":316},[265,37],[],[89,21],[318,319],"CrunchUP 食物识别","Gemini 营养估算","CrunchUP食物识别-5384","10.32628\u002Fcseit2612569",{"doi":321,"openalex_id":323,"authors":324,"venue":301,"cited_by_count":15,"oa_url":329,"card":330,"direction":293,"ingested_from":64},"W7222890719",[325,327],{"name":326,"orcid":9},"Vijayaraghavan R",{"name":328,"orcid":9},"V. S. Anita Sofia","https:\u002F\u002Fijsrcseit.com\u002Fhome\u002Farticle\u002Fdownload\u002FCSEIT2612569\u002FCSEIT2612569",{"tldr":331,"method":332,"finding":333,"direction":136,"opportunity":334},"开发CrunchUP网页应用，拍照识别食物并估算营养，结合个人目标给出膳食建议。","FastAPI+React架构，Gemini 2.5 Flash视觉模型识别食物","视觉语言服务能降低饮食记录负担，但结果应作参考而非临床测量。","可研究校准份量参考、可信食物数据库与置信度报告，提升营养估算准确性。","2026-10-20T23:30:09.983165Z",{"id":337,"title":338,"url":339,"summary":340,"summary_zh":9,"content":9,"source_name":341,"source_url":9,"published_at":342,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":343,"score_detail":344,"sources":347,"tags":349,"primary_topic":352,"secondary_topics":358,"classification":360,"search_phrases":376,"slug":379,"view_count":15,"doi":9,"paper":380,"created_at":388},5363,"牛肉牛区域特征计算机视觉分析:基于 YOLO 与深度估计的牛体况自动评分","https:\u002F\u002Fbioengineer.org\u002Fcameras-learn-to-read-a-cows-spine-tail-head-and-body-shape-to-score-its-condition","Islam M.N., Schrick F.N., Edwards J.L., Burns R.T., Qi H., Gan H. (2026) 开发一种基于区域特定计算机视觉特征的牛肉牛体况自动评分(BCS)方法。研究通过 YOLO 检测 + 深度估计提取脊柱、尾根、体廓等区域特征,使用 SHAP 进行可解释 AI 分析,并以 XGBoost 融合。在 5 折交叉验证中两种模式均稳定。验证集仅 72 只动物且极端值样本少,深度图来自推断而非实测。框架可作为后续发展的方法学基础。DOI: 10.1016\u002Fj.atech.2026.102628。","Smart Agricultural Technology Vol.15 Article 102628 2026-10-15","2026-10-15T00:00:00Z",68,{"impact":83,"substance":345,"depth":82,"authority":148,"freshness":67,"relevant":68,"comment":346},20,"方法学上有新意，但验证样本仅72头且深度图为推断值，属细分领域论文进展，未达每日精选门槛。",[348],{"name":341,"url":339},[88,89,21,350,351],"肉牛养殖","体况评分",{"code":353,"slug":354,"name":355,"path":356,"parent_code":357},"C2","ruminants","草食牲畜","\u002Ftopic\u002Flivestock\u002Fruminants","C",[359],{"code":32,"slug":93,"name":94,"path":95,"parent_code":96},{"status":99,"locked":14,"version":27,"method":28,"evidence":361},{"candidates":362,"agricultural_evidence":34},[363,367,372],{"code":32,"score":103,"supported":34,"title":364,"abstract":365,"tags":366},[21,214],[21,214],[89,21],{"code":353,"score":83,"supported":34,"title":368,"abstract":370,"tags":371},[369],"肉牛",[369],[350],{"code":109,"score":68,"supported":14,"title":373,"abstract":374,"tags":375},[],[],[88],[377,378],"YOLO 深度估计 牛体况评分","牛肉牛 计算机视觉 BCS","YOLO深度估计牛体况评分-5363",{"doi":9,"openalex_id":9,"authors":381,"venue":9,"cited_by_count":15,"oa_url":9,"card":382,"direction":63,"ingested_from":387},[],{"tldr":383,"method":384,"finding":385,"direction":63,"opportunity":386},"用YOLO检测与深度估计提取牛体区域特征，实现肉牛体况自动评分。","YOLO检测、深度估计、SHAP可解释分析与XGBoost融合，5折交叉验证。","区域特征融合评分稳定，但验证集仅72头且深度图为推断值。","可引入实测深度传感器与更大样本，提升极端体况评分鲁棒性。","agent","2026-10-20T02:47:54.361149Z"]