[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-3670":3,"related-3670":79},{"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":42,"slug":45,"view_count":46,"doi":47,"paper":48,"created_at":78},3670,"Automated pavement marking retroreflectivity estimation via computer vision and machine learning","https:\u002F\u002Fdoi.org\u002F10.1016\u002Fj.trip.2026.102283","Pavement marking retroreflectivity is essential for nighttime road safety but deteriorates due to environmental and traffic factors. Traditional assessment methods are costly and labor-intensive. This paper introduces a sensor-independent, artificial intelligence driven methodology for predicting pavement-marking retroreflectivity using just Red, Green, and Blue (RGB) smartphone videos. The artificial intelligence contribution is through combining gray level co-occurrence matrix (GLCM) based texture and contrast features with a physics-informed synthetic labeling model that imitates material deterioration and environmental influences, allowing for supervised learning in the absence of specialized sensors. Five machine learning models were trained using eight features derived from video-based image processing and physics-informed synthetic labels. XGBoost achieved the highest internal consistency with the synthetic labeling model (R 2 = 0.98 for yellow and 0.97 for white), demonstrating that machine learning algorithms can effectively approximate complex, physics-based relationships between retroreflectivity and standard RGB video features. These results establish proof-of-concept feasibility; field validation against ground-truth retroreflectivity measurements remains a necessary next step before operational deployment. Future work will validate the model using field measurements and expand the dataset across varied pavement types, weather conditions, and traffic environments to enhance generalizability.","路面标线逆反射性能对夜间道路安全至关重要，但会因环境和交通因素而衰减。传统评估方法成本高且劳动强度大。本文提出了一种与传感器无关、由人工智能驱动的方法，仅使用红、绿、蓝（RGB）智能手机视频即可预测路面标线逆反射性能。其人工智能贡献在于将基于灰度共生矩阵（GLCM）的纹理与对比度特征，与一个模拟材料劣化和环境影响的物理信息合成标注模型相结合，从而在缺乏专用传感器的情况下实现监督学习。研究使用由基于视频的图像处理和物理信息合成标签得到的八个特征训练了五种机器学习模型。XGBoost在与合成标注模型的一致性方面表现最佳（黄色R²=0.98，白色R²=0.97），表明机器学习算法能够有效逼近逆反射性能与标准RGB视频特征之间复杂的、基于物理的关系。这些结果确立了概念验证的可行性；在投入实际应用之前，仍需针对真实逆反射测量值开展现场验证。未来工作将利用现场测量验证模型，并扩展数据集以涵盖不同路面类型、天气条件和交通环境，从而增强泛化能力。",null,"Transportation Research Interdisciplinary Perspectives","2026-09-27T00:00:00Z","论文",10,false,0,{"impact":15,"substance":15,"depth":15,"authority":15,"freshness":15,"relevant":15,"comment":17},"该论文研究路面标线逆反射性能的计算机视觉估算方法，属于道路交通安全与交通基础设施领域，与三农、农业信息化、智慧农业、数字乡村等主题无关，不建议进入每日精选。",[19],{"name":10,"url":6},[21,22,23],"计算机视觉","机器学习","道路安全",[],{"status":26,"locked":14,"version":27,"method":28,"evidence":29},"needs_review","2026-10-22.2","rules",{"candidates":30,"agricultural_evidence":14,"reason":41},[31],{"code":32,"score":33,"supported":34,"title":35,"abstract":38,"tags":40},"I2",27,true,[36,37],"machine learning","computer vision",[22,36,39],"artificial intelligence",[22,21],"农业相关性需要复核",[43,44],"计算机视觉 机器学习 道路安全","计算机视觉 机器学习","计算机视觉机器学习道路安全-3670",1,"10.1016\u002Fj.trip.2026.102283",{"doi":47,"openalex_id":49,"authors":50,"venue":10,"cited_by_count":15,"oa_url":69,"card":70,"direction":76,"ingested_from":77},"W7214556100",[51,54,56,59,62,64,67],{"name":52,"orcid":53},"Mark Gervas Ngotonie","https:\u002F\u002Forcid.org\u002F0009-0003-2558-9034",{"name":55,"orcid":9},"Nana Kankam Gyimah",{"name":57,"orcid":58},"Gurcan Comert","https:\u002F\u002Forcid.org\u002F0000-0002-2373-5013",{"name":60,"orcid":61},"Judith Mwakalonge","https:\u002F\u002Forcid.org\u002F0000-0002-7497-6829",{"name":63,"orcid":9},"Said Siuhi",{"name":65,"orcid":66},"Akinbobola Jegede","https:\u002F\u002Forcid.org\u002F0009-0002-4015-6168",{"name":68,"orcid":9},"Ayobami Taiwo","https:\u002F\u002Fwww.sciencedirect.com\u002Fscience\u002Farticle\u002Fpii\u002FS2590198226004483\u002Fpdf",{"tldr":71,"method":72,"finding":73,"direction":74,"opportunity":75},"用手机RGB视频和机器学习预测路面标线逆反射率，无需专用传感器。","GLCM纹理特征+物理信息合成标签，训练五种机器学习模型。","XGBoost与合成标签模型一致性最高（R²=0.98黄\u002F0.97白），验证概念可行。","其他","需实地验证并扩展多路面、天气和交通场景数据集，提升模型泛化能力。","农业遥感与作物表型","openalex","2026-09-28T23:30:27.891115Z",{"total":80,"page":46,"page_size":80,"items":81},6,[82,150,204,298,341,411],{"id":83,"title":84,"url":85,"summary":86,"summary_zh":87,"content":9,"source_name":88,"source_url":85,"published_at":89,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":90,"score_detail":91,"sources":97,"tags":99,"primary_topic":104,"secondary_topics":109,"classification":110,"search_phrases":129,"slug":132,"view_count":15,"doi":133,"paper":134,"created_at":149},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":92,"substance":93,"depth":94,"authority":95,"freshness":13,"relevant":46,"comment":96},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。",[98],{"name":88,"url":85},[100,101,21,102,103],"智慧农业","农业人工智能","产量预测","传感器校准",{"code":32,"slug":105,"name":106,"path":107,"parent_code":108},"agricultural-ai","农业AI与大模型","\u002Ftopic\u002Fdigital-agriculture\u002Fagricultural-ai","I",[],{"status":111,"locked":14,"version":27,"method":28,"evidence":112},"classified",{"candidates":113,"agricultural_evidence":34},[114,119,124],{"code":32,"score":115,"supported":34,"title":116,"abstract":117,"tags":118},24,[39,37],[21,37],[101,21],{"code":120,"score":46,"supported":14,"title":121,"abstract":122,"tags":123},"I1",[],[],[100],{"code":125,"score":46,"supported":14,"title":126,"abstract":127,"tags":128},"I3",[],[],[102],[130,131],"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":133,"openalex_id":135,"authors":136,"venue":88,"cited_by_count":15,"oa_url":85,"card":143,"direction":147,"ingested_from":77},"W7223104380",[137,139,141],{"name":138,"orcid":9},"Tigist Alemu Gebremariam",{"name":140,"orcid":9},"Hawi Tesfaye Desta",{"name":142,"orcid":9},"Mulugeta Bekele Wolde",{"tldr":144,"method":145,"finding":146,"direction":147,"opportunity":148},"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":151,"title":152,"url":153,"summary":154,"summary_zh":155,"content":9,"source_name":88,"source_url":153,"published_at":89,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":156,"score_detail":157,"sources":161,"tags":163,"primary_topic":166,"secondary_topics":167,"classification":168,"search_phrases":184,"slug":187,"view_count":15,"doi":188,"paper":189,"created_at":203},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":95,"substance":158,"depth":93,"authority":159,"freshness":13,"relevant":46,"comment":160},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。",[162],{"name":88,"url":153},[100,101,21,164,165],"作物病害","气候变化",{"code":32,"slug":105,"name":106,"path":107,"parent_code":108},[],{"status":111,"locked":14,"version":27,"method":28,"evidence":169},{"candidates":170,"agricultural_evidence":34},[171,175,180],{"code":32,"score":115,"supported":34,"title":172,"abstract":173,"tags":174},[39,37],[21,37],[101,21],{"code":176,"score":46,"supported":14,"title":177,"abstract":178,"tags":179},"H7",[],[],[165],{"code":120,"score":46,"supported":14,"title":181,"abstract":182,"tags":183},[],[],[100],[185,186],"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":188,"openalex_id":190,"authors":191,"venue":88,"cited_by_count":15,"oa_url":153,"card":198,"direction":147,"ingested_from":77},"W7223087892",[192,194,196],{"name":193,"orcid":9},"Kossi Afeli",{"name":195,"orcid":9},"Akouvi Amegashie",{"name":197,"orcid":9},"Komlan Sélom Dossou",{"tldr":199,"method":200,"finding":201,"direction":147,"opportunity":202},"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":205,"title":206,"url":207,"summary":208,"summary_zh":209,"content":9,"source_name":210,"source_url":207,"published_at":89,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":211,"score_detail":212,"sources":216,"tags":218,"primary_topic":222,"secondary_topics":228,"classification":236,"search_phrases":274,"slug":277,"view_count":15,"doi":278,"paper":279,"created_at":297},5558,"Estimation of Soil Water Content in Winter Wheat Field at Jointing Stage Using Multi-Source Features and Machine Learning","https:\u002F\u002Fdoi.org\u002F10.3390\u002Fagronomy16202071","Accurate and timely monitoring of soil water content (SWC) plays an essential supporting role in crop irrigation management and smart agricultural practice. Remote sensing technology provides rapid and convenient approaches to large-scale farmland SWC estimation or inversion. However, existing mechanistic modeling methods are limited by parameter localization and land surface heterogeneity, while machine learning data-driven methods lack quite good physical interpretability. To address deficiencies in single data-driven methods and single vegetation microwave scattering models, and explore the estimation effect of different feature variables in machine learning. This study proposes an innovative estimation model that integrates machine learning strategies and multi-source features, including spectral indices and features derived from the classic water cloud model (WCM). The novelty of the proposed model lies in balancing physical interpretability and data-driven accuracy, namely extracting relevant soil backscattering parameters through WCM, which are then used as input features in the machine learning. A combination of field-measured data and unmanned aerial vehicle (UAV) multi-source remote sensing data, including multispectral and synthetic aperture radar (SAR) data, was acquired synchronously at the jointing stage of winter wheat. Through extracting variables such as the SAR microwave backscattering coefficient and the optical vegetation index, and conducting feature screening and parameter sensitivity analysis, we optimized the input variable sets of models based on back-propagation (BP) neural network, support vector regression (SVR) and random forest (RF) algorithms. The study results showed that the established estimation model had excellent estimation performance, especially for the RF algorithm, with the highest R2 of 0.855, lowest MAE of 0.0212 cm3\u002Fcm3, and lowest RMSE of 0.0253 cm3\u002Fcm3. It was also shown that different feature input schemes have obvious effects on SWC estimation performance. On the whole, RF had the best estimation accuracy, followed by SVR and then BPNN. This study demonstrated that models integrating multi-source remote sensing data and machine learning could enhance SWC estimation ability for winter wheat fields, and could also serve as a promising approach for SWC monitoring in winter wheat fields in the North China Plain.","准确、及时地监测土壤含水量（SWC）对作物灌溉管理和智慧农业实践具有重要的支撑作用。遥感技术为大规模农田土壤含水量估算或反演提供了快速便捷的手段。然而，现有机理建模方法受参数本地化和地表异质性限制，机器学习数据驱动方法则缺乏良好的物理解释性。针对单一数据驱动方法和单一植被微波散射模型的不足，并探究不同特征变量在机器学习中的估算效果，本研究提出了一种融合机器学习策略与多源特征的创新估算模型，所采用特征包括光谱指数和经典水云模型（WCM）衍生的特征。该模型的新颖性在于平衡物理解释性与数据驱动精度，即通过WCM提取相关土壤后向散射参数，再将其作为机器学习的输入特征。研究在冬小麦拔节期同步获取了田间实测数据与无人机（UAV）多源遥感数据，包括多光谱和合成孔径雷达（SAR）数据。通过提取SAR微波后向散射系数和光学植被指数等变量，并进行特征筛选和参数敏感性分析，优化了基于反向传播（BP）神经网络、支持向量回归（SVR）和随机森林（RF）算法的模型输入变量集。研究结果表明，所建立的估算模型具有优良的估算性能，尤其是RF算法，最高R²为0.855，最低MAE为0.0212 cm³\u002Fcm³，最低RMSE为0.0253 cm³\u002Fcm³。研究还表明，不同特征输入方案对土壤含水量估算性能有明显影响。总体而言，RF估算精度最佳，SVR次之，BPNN再次之。本研究证明，融合多源遥感数据与机器学习的模型能够提升冬小麦田土壤含水量估算能力，也可为华北平原冬小麦田土壤含水量监测提供一种有前景的方法。","Agronomy",77,{"impact":213,"substance":214,"depth":94,"authority":159,"freshness":13,"relevant":46,"comment":215},16,21,"该研究融合多源遥感与机器学习估算冬小麦拔节期土壤含水量，方法有创新、结论可靠，对智慧灌溉有参考价值，但属细分领域论文，影响范围有限。",[217],{"name":210,"url":207},[100,22,219,220,221],"遥感监测","土壤墒情","冬小麦",{"code":223,"slug":224,"name":225,"path":226,"parent_code":227},"A2","wheat","小麦","\u002Ftopic\u002Ffield-crops\u002Fwheat","A",[229,235],{"code":230,"slug":231,"name":232,"path":233,"parent_code":234},"H2","soil","土壤健康与改良","\u002Ftopic\u002Fenvironment\u002Fsoil","H",{"code":32,"slug":105,"name":106,"path":107,"parent_code":108},{"status":111,"locked":14,"version":27,"method":28,"evidence":237},{"candidates":238,"agricultural_evidence":34},[239,245,251,255,262,269],{"code":223,"score":240,"supported":34,"title":241,"abstract":243,"tags":244},26,[242,224],"winter wheat",[225,242,224],[221],{"code":230,"score":246,"supported":34,"title":247,"abstract":248,"tags":250},15,[231],[249,231],"土壤",[220],{"code":32,"score":246,"supported":34,"title":252,"abstract":253,"tags":254},[36],[22,36],[22],{"code":125,"score":13,"supported":34,"title":256,"abstract":257,"tags":261},[],[258,259,260],"遥感","remote sensing","multispectral",[219],{"code":263,"score":80,"supported":34,"title":264,"abstract":265,"tags":268},"H3",[],[266,267],"灌溉","irrigation",[],{"code":120,"score":270,"supported":34,"title":271,"abstract":272,"tags":273},4,[],[100],[100],[275,276],"冬小麦 土壤含水量 遥感","多源遥感 机器学习 墒情","冬小麦土壤含水量遥感-5558","10.3390\u002Fagronomy16202071",{"doi":278,"openalex_id":280,"authors":281,"venue":210,"cited_by_count":15,"oa_url":207,"card":291,"direction":296,"ingested_from":77},"W7223156204",[282,285,287,289],{"name":283,"orcid":284},"Mingliang Jiang","https:\u002F\u002Forcid.org\u002F0000-0002-6486-6087",{"name":286,"orcid":9},"Yi Zhang",{"name":288,"orcid":9},"WEI Lei",{"name":290,"orcid":9},"Chengcai Zhang",{"tldr":292,"method":293,"finding":294,"direction":76,"opportunity":295},"融合WCM物理特征与多源遥感，用机器学习估算冬小麦拔节期土壤含水量。","无人机多光谱与SAR数据，WCM参数+光谱指数，BP\u002FSVR\u002FRF建模。","RF最优，R²达0.855，RMSE为0.0253 cm³\u002Fcm³，特征方案影响显著。","可探索物理约束与深度学习耦合，并迁移至多生育期、多作物及不同土壤质地验证。","智慧农业 \u002F 农业物联网","2026-10-22T11:20:18.887301Z",{"id":299,"title":300,"url":301,"summary":302,"summary_zh":303,"content":9,"source_name":304,"source_url":301,"published_at":305,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":15,"score_detail":306,"sources":308,"tags":310,"primary_topic":9,"secondary_topics":312,"classification":313,"search_phrases":322,"slug":325,"view_count":15,"doi":326,"paper":327,"created_at":340},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":307},"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。",[309],{"name":304,"url":301},[101,21,311],"机器人感知",[],{"status":26,"locked":14,"version":27,"method":28,"evidence":314},{"candidates":315,"agricultural_evidence":14,"reason":41},[316],{"code":32,"score":33,"supported":34,"title":317,"abstract":319,"tags":321},[37,318],"YOLO",[320,21,318],"目标检测",[101,21],[323,324],"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":326,"openalex_id":328,"authors":329,"venue":304,"cited_by_count":15,"oa_url":334,"card":335,"direction":147,"ingested_from":77},"W7222933754",[330,332],{"name":331,"orcid":9},"Aafareen Sikandar Shaikh",{"name":333,"orcid":9},"Archana Nilesh Pawar","https:\u002F\u002Fwww.irejournals.com\u002Fformatedpaper\u002F1724117.pdf",{"tldr":336,"method":337,"finding":338,"direction":147,"opportunity":339},"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":342,"title":343,"url":344,"summary":345,"summary_zh":346,"content":9,"source_name":347,"source_url":344,"published_at":348,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":15,"score_detail":349,"sources":351,"tags":353,"primary_topic":356,"secondary_topics":357,"classification":363,"search_phrases":388,"slug":391,"view_count":15,"doi":392,"paper":393,"created_at":410},5485,"Integrating machine learning and geospatial technologies for spatio-temporal analysis of urban expansion and land use transformation in Gambella Town, Ethiopia","https:\u002F\u002Fdoi.org\u002F10.1371\u002Fjournal.pone.0360419","Rapid urbanization and the associated expansion of built-up areas are transforming land resources, ecosystem functions, and the spatial structure of cities. Distinguishing urban growth, urban expansion, and urban sprawl is essential for understanding the processes and patterns of urban transformation and supporting sustainable urban planning. This study integrates machine learning and geospatial technologies to analyze spatio-temporal land use and land cover (LULC) dynamics and urban expansion in Gambella Town, Ethiopia. Multi-temporal Landsat satellite images from 2005, 2015, and 2025 were processed using remote sensing techniques and a supervised machine learning approach based on the Random Forest (RF) classifier. Six major LULC classes were identified: Agricultural Land, Bush with Grass, Built-Up Area, Bare Land, Forest Land, and Water Bodies. The classification achieved overall accuracies of 90%, 90%, and 89% for 2005, 2015, and 2025, respectively. The results revealed substantial land transformation and rapid physical urban expansion, with the built-up area increasing from 93 ha (2.0%) in 2005–1,956 ha (41.3%) in 2025. This expansion was accompanied by major transformations of surrounding land resources, particularly the conversion of Bush with Grass and other open areas to built-up land. Agricultural land increased from 634 ha (13.4%) to 1,750 ha (36.9%), whereas Bush with Grass declined from 3,375 ha (71.1%) to 652 ha (13.7%). Socio-economic information collected from 277 respondents indicated that population growth, rural–urban migration, housing demand, employment opportunities, and infrastructure development were major drivers of urban expansion. The study distinguishes urban expansion, represented by the increase in built-up land, from urban sprawl, which refers to the spatial pattern and form of outward urban development. The integration of Random Forest classification and GIS-based spatial analysis provides valuable evidence of land transformation, vegetation loss, and increasing pressure on urban services, supporting more sustainable urban planning and land management in Gambella Town.","快速城市化及与之相关的建成区扩张正在改变土地资源、生态系统功能和城市空间结构。区分城市增长（urban growth）、城市扩张（urban expansion）和城市蔓延（urban sprawl），对于理解城市转型的过程与格局、支撑可持续城市规划至关重要。本研究融合机器学习与地理空间技术，分析埃塞俄比亚甘贝拉镇（Gambella Town）土地利用与土地覆盖（LULC）的时空动态及城市扩张。采用遥感技术和基于随机森林（RF）分类器的监督机器学习方法，处理了2005年、2015年和2025年的多时相Landsat卫星影像。研究识别出六类主要LULC类型：农业用地、灌草地、建成区、裸地、林地和水体。2005年、2015年和2025年的分类总体精度分别为90%、90%和89%。结果表明，土地发生了显著转变，城市实体空间快速扩张，建成区面积从2005年的93公顷（2.0%）增至2025年的1,956公顷（41.3%）。这一扩张伴随着周边土地资源的重大转变，尤其是灌草地及其他开阔地被转化为建成用地。农业用地从634公顷（13.4%）增至1,750公顷（36.9%），而灌草地则从3,375公顷（71.1%）降至652公顷（13.7%）。从277名受访者收集的社会经济信息表明，人口增长、城乡迁移、住房需求、就业机会和基础设施建设是城市扩张的主要驱动因素。本研究将城市扩张（以建成用地增加为表征）与城市蔓延（指城市向外发展的空间格局与形态）加以区分。随机森林分类与基于GIS的空间分析的融合，为土地转变、植被丧失及城市服务压力加剧提供了有价值的证据，有助于甘贝拉镇实现更可持续的城市规划与土地管理。","PLoS ONE","2026-10-20T00:00:00Z",{"impact":15,"substance":15,"depth":15,"authority":15,"freshness":15,"relevant":15,"comment":350},"研究聚焦埃塞俄比亚城市扩张与土地利用变化，属城市地理与遥感应用，与三农、农业信息化主题无直接关联，不建议进入每日精选。",[352],{"name":347,"url":344},[22,219,354,355],"土地利用","城市扩张",{"code":32,"slug":105,"name":106,"path":107,"parent_code":108},[358],{"code":359,"slug":360,"name":361,"path":362,"parent_code":108},"I4","gis","GIS与空间分析","\u002Ftopic\u002Fdigital-agriculture\u002Fgis",{"status":111,"locked":14,"version":27,"method":28,"evidence":364},{"candidates":365,"agricultural_evidence":34},[366,370,378,383],{"code":32,"score":246,"supported":34,"title":367,"abstract":368,"tags":369},[36],[22,36],[22],{"code":359,"score":371,"supported":34,"title":372,"abstract":373,"tags":377},9,[],[374,375,376],"空间分析","GIS","spatial analysis",[],{"code":125,"score":379,"supported":34,"title":380,"abstract":381,"tags":382},7,[],[258,259],[219],{"code":384,"score":46,"supported":14,"title":385,"abstract":386,"tags":387},"H6",[],[],[354],[389,390],"Gambella Town 城市扩张","Random Forest 土地利用分类","GambellaTown城市扩张-5485","10.1371\u002Fjournal.pone.0360419",{"doi":392,"openalex_id":394,"authors":395,"venue":347,"cited_by_count":15,"oa_url":344,"card":405,"direction":76,"ingested_from":77},"W7223022155",[396,399,401,403],{"name":397,"orcid":398},"Kanenus Fufa Dararo","https:\u002F\u002Forcid.org\u002F0009-0007-5747-1558",{"name":400,"orcid":9},"Goanar Jal Dup",{"name":402,"orcid":9},"Weyesa Merga Chemeda",{"name":404,"orcid":9},"Shankar Karuppannan",{"tldr":406,"method":407,"finding":408,"direction":76,"opportunity":409},"用机器学习与遥感分析埃塞俄比亚甘贝拉镇2005-2025年城市扩张与土地利用变化。","Landsat多时相影像、随机森林分类与GIS空间分析，结合277份问卷。","建成区从93公顷增至1956公顷，灌草地大幅转为建成区和农地。","可将随机森林时序分类与驱动力调查结合，扩展到更多非洲小城镇并耦合城市增长模拟。","2026-10-21T23:30:23.271293Z",{"id":412,"title":413,"url":414,"summary":415,"summary_zh":9,"content":9,"source_name":416,"source_url":9,"published_at":417,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":418,"score_detail":419,"sources":422,"tags":424,"primary_topic":428,"secondary_topics":429,"classification":430,"search_phrases":450,"slug":453,"view_count":15,"doi":9,"paper":454,"created_at":462},5462,"Architecture diverse TabNet ensemble for crop type classification using harmonized multisource soil data（基于多源土壤数据融合的架构多样性TabNet集成作物类型分类）","https:\u002F\u002Fwww.nature.com\u002Farticles\u002Fs41598-026-75366-7","对18,237条多源土壤记录、7类作物构建TabNet集成框架。TPE优化使对数损失由0.3325降至0.2566，集成后准确率0.9053、宏F1 0.9025，优于XGBoost的0.8724。作者称尚未做外部验证。","Scientific Reports 2026-10-18","2026-10-18T00:00:00Z",70,{"impact":95,"substance":420,"depth":94,"authority":159,"freshness":92,"relevant":46,"comment":421},20,"方法新颖、数据规模可观且指标有实质提升，但缺乏外部验证、属细分方法研究，公共影响有限，可作候选而非必选。",[423],{"name":416,"url":414},[101,22,425,426,427],"精准农业","土壤数据","作物分类",{"code":230,"slug":231,"name":232,"path":233,"parent_code":234},[],{"status":111,"locked":14,"version":27,"method":28,"evidence":431},{"candidates":432,"agricultural_evidence":34},[433,437,442,446],{"code":230,"score":420,"supported":34,"title":434,"abstract":435,"tags":436},[249,231],[249],[426],{"code":32,"score":438,"supported":14,"title":439,"abstract":440,"tags":441},2,[],[],[101,22],{"code":120,"score":46,"supported":14,"title":443,"abstract":444,"tags":445},[],[],[425],{"code":125,"score":46,"supported":14,"title":447,"abstract":448,"tags":449},[],[],[427],[451,452],"TabNet 作物类型分类","多源土壤数据 作物分类","TabNet作物类型分类-5462",{"doi":9,"openalex_id":9,"authors":455,"venue":9,"cited_by_count":15,"oa_url":9,"card":456,"direction":147,"ingested_from":461},[],{"tldr":457,"method":458,"finding":459,"direction":147,"opportunity":460},"构建架构多样性TabNet集成，用多源土壤数据分类7类作物。","18,237条多源土壤记录，TPE调参的TabNet集成，对比XGBoost。","集成后准确率0.9053、宏F1 0.9025，优于XGBoost的0.8724。","尚未外部验证，可跨区域迁移验证并融合遥感时序特征提升泛化。","agent","2026-10-21T00:10:08.428584Z"]