[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-1310":3,"related-1310":97},{"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":22,"tags":24,"primary_topic":30,"secondary_topics":36,"classification":37,"search_phrases":65,"slug":68,"view_count":69,"doi":70,"paper":71,"created_at":96},1310,"Object Detection in Computer Vision: A Comprehensive Review of Algorithms, Datasets, and Evaluation Techniques","https:\u002F\u002Fdoi.org\u002F10.20944\u002Fpreprints202608.2288.v1","Object detection has evolved dramatically over three decades, transitioning from handcrafted features to deep learning, transformers, and now large language models. This paper provides a comprehensive systematic review of this evolution, covering traditional methods (SIFT, HOG, Viola-Jones), CNN-based single- and two-stage detectors (R-CNN family, YOLO v1–v12, SSD), transformer-based architectures (DETR, ViT, Swin), and emerging LLM-driven approaches (GPT-4V, Grounding DINO, YOLO-World). We benchmark and compare these methods using key metrics—mAP, IoU, FPS—and analyze over 50 datasets across generic, autonomous driving, medical, remote sensing, and agricultural domains. Our critical synthesis identifies persistent research gaps: lightweight edge deployment, sensor fusion for real-time systems, bias and fairness in AI, and multimodal vision-language integration. By consolidating three decades of progress and outlining future challenges, this survey serves as a valuable resource for researchers and practitioners developing accurate, efficient, and reliable object detection systems.","目标检测在过去三十年间经历了深刻的演变，从手工特征工程发展到深度学习、Transformer架构，直至如今的大语言模型。本文对这一演进过程进行了全面的系统性综述，涵盖了传统方法（SIFT、HOG、Viola-Jones）、基于CNN的单阶段和双阶段检测器（R-CNN系列、YOLO v1–v12、SSD）、基于Transformer的架构（DETR、ViT、Swin），以及新兴的由大语言模型驱动的方法（GPT-4V、Grounding DINO、YOLO-World）。我们使用关键指标——mAP、IoU、FPS——对这些方法进行了基准测试和比较，并分析了涵盖通用、自动驾驶、医学、遥感及农业领域的50多个数据集。我们的批判性综合指出了持续存在的研究空白：轻量级边缘部署、用于实时系统的传感器融合、人工智能中的偏见与公平性，以及多模态视觉-语言集成。通过整合三十年的进展并勾勒未来挑战，本综述为开发准确、高效且可靠的目标检测系统的研究人员和从业者提供了宝贵的资源。",null,"Preprints.org","2026-08-31T00:00:00Z","论文",10,false,56,{"impact":17,"substance":18,"depth":18,"authority":13,"freshness":19,"relevant":20,"comment":21},8,18,2,1,"综述覆盖农业领域目标检测应用，但时效性低，影响有限。",[23],{"name":10,"url":6},[25,26,27,28,29],"智慧农业","农业人工智能","综述","深度学习","目标检测",{"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":14,"version":39,"method":40,"evidence":41},"classified","2026-10-22.2","rules",{"candidates":42,"agricultural_evidence":45},[43,52,60],{"code":31,"score":44,"supported":45,"title":46,"abstract":48,"tags":51},20,true,[47],"computer vision",[29,28,49,50],"deep learning","YOLO",[26,28,29],{"code":53,"score":54,"supported":45,"title":55,"abstract":56,"tags":59},"I3",6,[],[57,58],"遥感","remote sensing",[],{"code":61,"score":20,"supported":14,"title":62,"abstract":63,"tags":64},"I1",[],[],[25],[66,67],"农业人工智能 智慧农业 深度学习 目标检测","农业人工智能 智慧农业","农业人工智能智慧农业深度学习目标检测-1310",3,"10.20944\u002Fpreprints202608.2288.v1",{"doi":70,"openalex_id":72,"authors":73,"venue":10,"cited_by_count":87,"oa_url":6,"card":88,"direction":94,"ingested_from":95},"W7204770977",[74,77,79,81,83,85],{"name":75,"orcid":76},"Md. Faishal Rahaman","https:\u002F\u002Forcid.org\u002F0000-0001-5656-1542",{"name":78,"orcid":9},"Li Xueyuan",{"name":80,"orcid":9},"Md Nagib Mahfuz Sunny",{"name":82,"orcid":9},"Muhammad Amjad",{"name":84,"orcid":9},"Xin Gao",{"name":86,"orcid":9},"Sakib Hasan",0,{"tldr":89,"method":90,"finding":91,"direction":92,"opportunity":93},"系统综述目标检测三十年发展，从传统方法到深度学习、Transformer及大语言模型，并分析数据集与","系统综述，涵盖传统、CNN、Transformer及LLM方法，比较mAP、Io","识别出轻量边缘部署、传感器融合、公平性及多模态视觉语言集成等研究空白。","农业人工智能与决策模型","农业领域目标检测可探索轻量模型与多模态融合，结合遥感数据提升复杂场景下的实时检测精度。","农业遥感与作物表型","openalex","2026-09-01T23:30:19.994895Z",{"total":54,"page":20,"page_size":54,"items":98},[99,161,225,287,339,420],{"id":100,"title":101,"url":102,"summary":103,"summary_zh":9,"content":9,"source_name":104,"source_url":9,"published_at":105,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":106,"score_detail":107,"sources":113,"tags":115,"primary_topic":118,"secondary_topics":124,"classification":126,"search_phrases":148,"slug":151,"view_count":87,"doi":9,"paper":152,"created_at":160},5532,"精准棉花病害管理：融合无人机影像与定制YOLO深度学习框架","https:\u002F\u002Farxiv.org\u002Fabs\u002F2610.24182","研究提出单阶段全卷积检测器Cotton-Vision，融合幽灵瓶颈通道注意力与跨阶段部分融合三项轻量化设计，参数减少20%。在巴基斯坦旁遮普省7个棉田采集的2700余幅高分辨率无人机影像上测试，精确率97.7%、召回率77.7%、mAP@50为85.6%，较主流基线提升10至20个百分点。预印本，仅在巴基斯坦单一区域验证。","arXiv 预印本 2026-10-16","2026-10-16T00:00:00Z",69,{"impact":108,"substance":109,"depth":110,"authority":111,"freshness":54,"relevant":20,"comment":112},16,21,17,9,"方法轻量化与实测数据有增量，但仅单区域预印本验证，属细分技术进展，可入精选备选。",[114],{"name":104,"url":102},[25,26,29,116,117],"无人机遥感","棉花病害",{"code":119,"slug":120,"name":121,"path":122,"parent_code":123},"A6","cash-crops","棉麻糖与特色经作","\u002Ftopic\u002Ffield-crops\u002Fcash-crops","A",[125],{"code":31,"slug":32,"name":33,"path":34,"parent_code":35},{"status":38,"locked":14,"version":39,"method":40,"evidence":127},{"candidates":128,"agricultural_evidence":45},[129,133,140,144],{"code":31,"score":18,"supported":45,"title":130,"abstract":131,"tags":132},[28,50],[],[26,29],{"code":119,"score":134,"supported":45,"title":135,"abstract":137,"tags":139},11,[136],"棉花",[138],"cotton",[],{"code":61,"score":20,"supported":14,"title":141,"abstract":142,"tags":143},[],[],[25],{"code":53,"score":20,"supported":14,"title":145,"abstract":146,"tags":147},[],[],[116],[149,150],"Cotton-Vision 棉花病害","无人机影像 棉花 检测","Cotton-Vision棉花病害-5532",{"doi":9,"openalex_id":9,"authors":153,"venue":9,"cited_by_count":87,"oa_url":9,"card":154,"direction":94,"ingested_from":159},[],{"tldr":155,"method":156,"finding":157,"direction":94,"opportunity":158},"提出轻量化检测器Cotton-Vision，用无人机影像精准识别棉花病害。","基于YOLO改进，融合幽灵瓶颈通道注意力与跨阶段部分融合，用2700余幅无人机影","精确率97.7%、mAP@50达85.6%，较主流基线提升10至20个百分点。","仅在巴基斯坦单一区域验证，可扩展至多区域多品种，并探索边缘设备实时部署。","agent","2026-10-22T00:07:09.816790Z",{"id":162,"title":163,"url":164,"summary":165,"summary_zh":166,"content":9,"source_name":167,"source_url":164,"published_at":168,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":169,"score_detail":170,"sources":174,"tags":176,"primary_topic":177,"secondary_topics":178,"classification":179,"search_phrases":196,"slug":199,"view_count":87,"doi":200,"paper":201,"created_at":224},5393,"IMPLEMENTATION OF META-HEURISTIC OPTIMIZATION ALGORITHMS IN CONVOLUTIONAL NEURAL NETWORKS FOR OBJECT DETECTION AND CLASSIFICATION","https:\u002F\u002Fdoi.org\u002F10.22190\u002Ffuacr251214006i","CNNs have grown to form the backbone of modern computer vision systems, producing outstanding object detection and classification performance within diverse application domains. However, the manual development of suitable CNN architectures and, correspondingly, the selection of optimal hyperparameter sets are still inconvenient and time-consuming processes. The aim of this survey is to contribute a comprehensive review regarding how meta-heuristic optimization algorithms such as Genetic Algorithms, Particle Swarm Optimization, and other swarm intelligence-based methods are embedded in the CNN-based pipelines with the intent to achieve better detection and classification results. We gather recent progress regarding automated architecture search, hyperparameter tuning, model compression, and ensemble learning. We identify how hybrid optimization approaches consistently yield superior accuracy with computational efficiency by systematically reviewing applications in medical imaging, agriculture, vehicular recognition, and remote sensing. Optimized CNN models outperform their baseline versions many times, especially in health diagnostics, where several fine-tuned architectures reach accuracies greater than 99%. This survey provides a structured and updated overview of methodological developments, implementation practices, and emerging research directions for those practitioners and researchers who will apply meta-heuristic optimization within deep learning-based computer vision systems.","卷积神经网络（CNNs）已发展成为现代计算机视觉系统的核心骨干，在多个应用领域中展现出卓越的目标检测与分类性能。然而，人工设计合适的CNN架构并相应地选择最优超参数集仍然是繁琐且耗时的过程。本综述旨在全面回顾遗传算法、粒子群优化及其他群体智能方法等元启发式优化算法如何嵌入基于CNN的流程中以获得更优的检测与分类结果。我们汇集了自动化架构搜索、超参数调优、模型压缩和集成学习方面的最新进展。通过系统梳理在医学影像、农业、车辆识别和遥感领域的应用，我们揭示了混合优化方法如何以计算效率持续实现更优的精度。优化后的CNN模型在多数情况下远超其基线版本，尤其在健康诊断领域，若干精细调优的架构达到了超过99%的准确率。本综述为将在基于深度学习的计算机视觉系统中应用元启发式优化的从业者和研究者，提供了关于方法学进展、实施实践和新兴研究方向的结构化且更新的概述。","Facta Universitatis Series Automatic Control and Robotics","2026-10-18T00:00:00Z",61,{"impact":17,"substance":18,"depth":171,"authority":172,"freshness":17,"relevant":20,"comment":173},15,12,"综述系统梳理元启发式优化在CNN视觉任务中的嵌入方法，含农业应用，但属通用方法综述，非农业专属突破。",[175],{"name":167,"url":164},[25,26,29,57],{"code":31,"slug":32,"name":33,"path":34,"parent_code":35},[],{"status":38,"locked":14,"version":39,"method":40,"evidence":180},{"candidates":181,"agricultural_evidence":45},[182,187,192],{"code":31,"score":134,"supported":45,"title":183,"abstract":184,"tags":186},[],[29,28,185,49,47],"计算机视觉",[26,29],{"code":53,"score":188,"supported":45,"title":189,"abstract":190,"tags":191},7,[],[57,58],[57],{"code":61,"score":20,"supported":14,"title":193,"abstract":194,"tags":195},[],[],[25],[197,198],"元启发式优化 卷积神经网络","CNN 目标检测","元启发式优化卷积神经网络-5393","10.22190\u002Ffuacr251214006i",{"doi":200,"openalex_id":202,"authors":203,"venue":167,"cited_by_count":87,"oa_url":164,"card":219,"direction":94,"ingested_from":95},"W7222920697",[204,207,210,213,216],{"name":205,"orcid":206},"Nikola Ivačko","https:\u002F\u002Forcid.org\u002F0000-0003-4367-119X",{"name":208,"orcid":209},"Ivan S Ciric","https:\u002F\u002Forcid.org\u002F0000-0003-0430-8937",{"name":211,"orcid":212},"Milica Tasić","https:\u002F\u002Forcid.org\u002F0009-0001-9160-6379",{"name":214,"orcid":215},"Ljiljana Radović","https:\u002F\u002Forcid.org\u002F0000-0002-4099-1289",{"name":217,"orcid":218},"Žarko Ćojbašić","https:\u002F\u002Forcid.org\u002F0000-0002-4581-1048",{"tldr":220,"method":221,"finding":222,"direction":92,"opportunity":223},"综述元启发式算法嵌入CNN流程以优化目标检测与分类的研究进展。","系统综述遗传算法、粒子群等群智能方法用于CNN架构搜索与超参调优。","混合优化方法在医学影像、农业等领域兼顾精度与效率，部分模型准确率超99%。","农业场景下元启发式优化CNN的架构搜索与轻量化部署仍缺乏系统基准与田间验证。","2026-10-20T23:30:19.233167Z",{"id":226,"title":227,"url":228,"summary":229,"summary_zh":230,"content":9,"source_name":231,"source_url":228,"published_at":168,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":232,"score_detail":233,"sources":235,"tags":237,"primary_topic":240,"secondary_topics":246,"classification":248,"search_phrases":266,"slug":269,"view_count":87,"doi":270,"paper":271,"created_at":286},5387,"LIGHTWEIGHT FEATURE FUSED ATTENTION DRIVEN MULTI LEVEL HYBRID MOBILENET V2 FRAMEWORK FOR EARLY TOMATO LEAF DISEASE DETECTION","https:\u002F\u002Fdoi.org\u002F10.62441\u002Factainnovations.vi.1918","It is crucial to find and treat tomato crop diseases early and accurately to improve crop efficiency and make sure there is no scarcity of crop. However, computerized analysis is provocative due to lighting differences, compound backgrounds, and graphic likenesses among infection regions. Research work proposed a lightweight Attention Driven Multi Level Hybrid MobileNet-V2 algorithm that fuses the handcrafted texture structures with Deep Learning (DL) features, then image preprocessing is incorporated using Gaussian, Bilateral, and Guided filters to minimize noise by preserving the disease lesions. LAB color model along with Contrast Limited Adaptive Histogram Equalization (CLAHE) are used to improve the contrast. Otsu thresholding is involved to effectively separate the leaf segment from diseased parts. Local Binary Pattern (LBP) and Gray Level Co-occurrence Matrix (GLCM) descriptors are used to obtain fine-grained features, while high level features are gained by pretrained MobileNet-V2 model. Features fusion is fed to the fully connected layers of deep learning model for training and validation purpose. Results have revealed that the suggested model attains better classification performance as compared to existing models, making it an effective explanation for real time tomato leaf disease detection for smart agriculture applications.","早期准确发现并治疗番茄作物病害，对于提高作物效率、确保作物不短缺至关重要。然而，由于光照差异、复杂背景以及感染区域之间的图形相似性，计算机化分析具有挑战性。研究提出了一种轻量级注意力驱动多层混合MobileNet-V2算法，该算法将手工纹理结构与深度学习（DL）特征相融合，然后结合高斯、双边和引导滤波器进行图像预处理，以在保留病害病变的同时最小化噪声。采用LAB颜色模型和对比度受限自适应直方图均衡化（CLAHE）来提高对比度。引入Otsu阈值分割以有效分离叶片片段与病害部分。使用局部二值模式（LBP）和灰度共生矩阵（GLCM）描述符获取细粒度特征，同时通过预训练的MobileNet-V2模型获取高层特征。特征融合后输入深度学习模型的全连接层进行训练和验证。结果表明，与现有模型相比，所提出的模型取得了更好的分类性能，使其成为智慧农业应用中实时番茄叶片病害检测的有效解决方案。","Acta Innovations",64,{"impact":172,"substance":18,"depth":108,"authority":13,"freshness":17,"relevant":20,"comment":234},"提出融合手工纹理特征与MobileNet-V2的轻量番茄叶病检测方法，有一定方法新颖性，但属单篇论文、应用范围有限，未达每日精选门槛。",[236],{"name":231,"url":228},[25,26,28,238,239],"番茄病害","图像识别",{"code":241,"slug":242,"name":243,"path":244,"parent_code":245},"B1","vegetables","蔬菜","\u002Ftopic\u002Fhorticulture\u002Fvegetables","B",[247],{"code":31,"slug":32,"name":33,"path":34,"parent_code":35},{"status":38,"locked":14,"version":39,"method":40,"evidence":249},{"candidates":250,"agricultural_evidence":45},[251,257,261],{"code":241,"score":171,"supported":45,"title":252,"abstract":254,"tags":256},[253],"tomato",[255,253],"番茄",[238],{"code":31,"score":111,"supported":45,"title":258,"abstract":259,"tags":260},[],[28,49],[26,239,28],{"code":61,"score":188,"supported":45,"title":262,"abstract":263,"tags":265},[],[25,264],"smart agriculture",[25],[267,268],"MobileNet-V2 番茄叶片病害","番茄叶病 深度学习 检测","MobileNet-V2番茄叶片病害-5387","10.62441\u002Factainnovations.vi.1918",{"doi":270,"openalex_id":272,"authors":273,"venue":231,"cited_by_count":87,"oa_url":279,"card":280,"direction":285,"ingested_from":95},"W7222977034",[274,276],{"name":275,"orcid":9},"Mrs. Vijetalaxmi A Hegde",{"name":277,"orcid":278},"Abhay Ramakant Kasetwar","https:\u002F\u002Forcid.org\u002F0000-0002-5800-8764","https:\u002F\u002Factainnovations.com\u002Findex.php\u002Fpub\u002Farticle\u002Fdownload\u002F1918\u002F981",{"tldr":281,"method":282,"finding":283,"direction":92,"opportunity":284},"提出轻量注意力多级混合MobileNet-V2框架，融合手工纹理与深度学习特征实现番茄叶片病害早期检","高斯\u002F双边\u002F引导滤波去噪，LAB+CLAHE增强，Otsu分割，LBP与GLCM","所提模型分类性能优于现有模型，可有效用于智慧农业实时番茄叶片病害检测。","可探索多作物跨域泛化、田间复杂光照下轻量模型部署与可解释性，及病害早期分级检测。","智慧农业 \u002F 农业物联网","2026-10-20T23:30:10.424511Z",{"id":288,"title":289,"url":290,"summary":291,"summary_zh":292,"content":9,"source_name":293,"source_url":290,"published_at":168,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":15,"score_detail":294,"sources":296,"tags":298,"primary_topic":300,"secondary_topics":301,"classification":302,"search_phrases":325,"slug":328,"view_count":87,"doi":329,"paper":330,"created_at":338},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",{"impact":17,"substance":18,"depth":108,"authority":54,"freshness":17,"relevant":20,"comment":295},"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。",[297],{"name":293,"url":290},[25,26,185,28,299],"水果品质检测",{"code":31,"slug":32,"name":33,"path":34,"parent_code":35},[],{"status":38,"locked":14,"version":39,"method":40,"evidence":303},{"candidates":304,"agricultural_evidence":45},[305,313,319],{"code":31,"score":306,"supported":45,"title":307,"abstract":308,"tags":312},28,[49,47],[29,28,309,185,310,49,311,47],"机器学习","machine learning","artificial intelligence",[26,28,185],{"code":61,"score":188,"supported":45,"title":314,"abstract":315,"tags":318},[],[316,317],"精准农业","precision agriculture",[25],{"code":320,"score":69,"supported":45,"title":321,"abstract":322,"tags":324},"B2",[],[323],"orchard",[],[326,327],"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":329,"openalex_id":331,"authors":332,"venue":293,"cited_by_count":87,"oa_url":290,"card":333,"direction":285,"ingested_from":95},"W7222892831",[],{"tldr":334,"method":335,"finding":336,"direction":94,"opportunity":337},"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。","2026-10-20T23:30:10.066042Z",{"id":340,"title":341,"url":342,"summary":343,"summary_zh":344,"content":9,"source_name":345,"source_url":342,"published_at":105,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":106,"score_detail":346,"sources":348,"tags":350,"primary_topic":353,"secondary_topics":358,"classification":360,"search_phrases":381,"slug":384,"view_count":20,"doi":385,"paper":386,"created_at":419},5164,"Lightweight Hybrid CNN-Transformer Network for Robust Maize Leaf Disease Detection and Edge Deployment","https:\u002F\u002Fdoi.org\u002F10.22266\u002Fijies2026.1130.50","Early and accurate detection of maize leaf diseases is essential for improving crop productivity and supporting precision agriculture.However, deep learning models trained on controlled leaf-image datasets often experience performance degradation under real-world agricultural conditions because of illumination variation, background clutter, occlusion, leaf orientation, and imaging noise.To address these challenges, this study develops a lightweight deep learning framework for robust maize leaf disease classification and edge-oriented deployment.The framework employs MobileNetV3-Small for efficient extraction of disease-discriminative visual features and investigates lightweight feature-learning strategies for improving classification under challenging field conditions.Experiments were conducted using PlantVillage-Maize and Field-Maize datasets comprising 6,188 images from four categories: Maize Leaf Blight (MLB), Common Rust, Healthy, and Leaf Spot.The image-only MobileNetV3-Small model achieved 97.8% accuracy and a 97.4% F1-score on the controlled PlantVillage-Maize dataset, while its performance decreased to 86.1% accuracy and an 83.9% F1-score on Field-Maize, highlighting the domain gap between controlled and real-field imagery.Robustness was further evaluated under Gaussian noise, illumination variation, and partial occlusion, together with computational-efficiency and edge-deployment analyses.The findings demonstrate the importance of lightweight and robust visual feature learning for practical maize disease diagnosis under real-world agricultural imaging conditions.","早期准确检测玉米叶片病害对于提高作物生产力和支持精准农业至关重要。然而，在受控叶片图像数据集上训练的深度学习模型，在真实农业条件下往往会出现性能下降，原因包括光照变化、背景杂乱、遮挡、叶片朝向以及成像噪声。为应对这些挑战，本研究开发了一个轻量级深度学习框架，用于稳健的玉米叶片病害分类和面向边缘的部署。该框架采用MobileNetV3-Small高效提取具有病害判别力的视觉特征，并研究轻量级特征学习策略，以提升在复杂田间条件下的分类性能。实验使用PlantVillage-Maize和Field-Maize数据集进行，共包含来自四个类别的6，188张图像：玉米大斑病（Maize Leaf Blight，MLB）、普通锈病（Common Rust）、健康（Healthy）和叶斑病（Leaf Spot）。仅使用图像的MobileNetV3-Small模型在受控的PlantVillage-Maize数据集上达到了97.8%的准确率和97.4%的F1分数，而在Field-Maize上的性能下降至86.1%的准确率和83.9%的F1分数，凸显了受控图像与真实田间图像之间的域差距。研究进一步在高斯噪声、光照变化和部分遮挡条件下评估了稳健性，并进行了计算效率和边缘部署分析。研究结果表明，轻量级且稳健的视觉特征学习对于真实农业成像条件下的实际玉米病害诊断具有重要意义。","International journal of intelligent engineering and systems",{"impact":172,"substance":44,"depth":108,"authority":172,"freshness":111,"relevant":20,"comment":347},"该研究提出轻量级玉米叶病检测框架，并揭示受控数据集与田间数据之间的性能差距，方法扎实、结论有参考价值，但属细分技术论文，公共影响有限。",[349],{"name":345,"url":342},[25,26,351,28,352],"边缘计算","玉米病害",{"code":354,"slug":355,"name":356,"path":357,"parent_code":123},"A3","maize","玉米","\u002Ftopic\u002Ffield-crops\u002Fmaize",[359],{"code":31,"slug":32,"name":33,"path":34,"parent_code":35},{"status":38,"locked":14,"version":39,"method":40,"evidence":361},{"candidates":362,"agricultural_evidence":45},[363,368,372,376],{"code":354,"score":364,"supported":45,"title":365,"abstract":366,"tags":367},14,[355],[356,355],[],{"code":31,"score":17,"supported":45,"title":369,"abstract":370,"tags":371},[],[28,49],[26,28],{"code":61,"score":188,"supported":45,"title":373,"abstract":374,"tags":375},[],[316,317],[25],{"code":377,"score":20,"supported":14,"title":378,"abstract":379,"tags":380},"I5",[],[],[351],[382,383],"MobileNetV3 玉米叶部病害","玉米叶片病害 边缘部署","MobileNetV3玉米叶部病害-5164","10.22266\u002Fijies2026.1130.50",{"doi":385,"openalex_id":387,"authors":388,"venue":345,"cited_by_count":87,"oa_url":342,"card":414,"direction":94,"ingested_from":95},"W7222602850",[389,392,395,397,399,402,404,406,409,411],{"name":390,"orcid":391},"Awwab Mohammad","https:\u002F\u002Forcid.org\u002F0000-0002-1148-5678",{"name":393,"orcid":394},"Punith kumar","https:\u002F\u002Forcid.org\u002F0009-0005-5676-4194",{"name":396,"orcid":9},"Bhavani K",{"name":398,"orcid":9},"Sailee N. Panchbhai",{"name":400,"orcid":401},"Sagarkumar S. Badhiye","https:\u002F\u002Forcid.org\u002F0000-0002-0710-3761",{"name":403,"orcid":9},"Manoj Kumar S B",{"name":405,"orcid":9},"Dharamvir",{"name":407,"orcid":408},"Sumit Kumar Sar","https:\u002F\u002Forcid.org\u002F0000-0003-0384-9088",{"name":410,"orcid":9},"Rani Chhagan Dalvi",{"name":412,"orcid":413},"Yogesh Harishchandra Bhosale","https:\u002F\u002Forcid.org\u002F0000-0001-6901-1419",{"tldr":415,"method":416,"finding":417,"direction":92,"opportunity":418},"构建轻量MobileNetV3-Small框架实现玉米叶片病害分类并评估边缘部署。","MobileNetV3-Small，PlantVillage-Maize与Fie","受控数据准确率97.8%，田间降至86.1%，揭示明显域差距。","可研究跨域自适应与田间鲁棒特征学习，缩小受控与真实场景性能差距。","2026-10-17T23:30:28.248539Z",{"id":421,"title":422,"url":423,"summary":424,"summary_zh":425,"content":9,"source_name":426,"source_url":423,"published_at":105,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":427,"score_detail":428,"sources":431,"tags":433,"primary_topic":436,"secondary_topics":437,"classification":439,"search_phrases":455,"slug":458,"view_count":20,"doi":459,"paper":460,"created_at":494},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",76,{"impact":108,"substance":109,"depth":110,"authority":429,"freshness":111,"relevant":20,"comment":430},13,"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。",[432],{"name":426,"url":423},[25,26,28,434,435],"病虫害识别","烟草",{"code":119,"slug":120,"name":121,"path":122,"parent_code":123},[438],{"code":31,"slug":32,"name":33,"path":34,"parent_code":35},{"status":38,"locked":14,"version":39,"method":40,"evidence":440},{"candidates":441,"agricultural_evidence":45},[442,447,451],{"code":119,"score":171,"supported":45,"title":443,"abstract":445,"tags":446},[444],"tobacco",[435,444],[435],{"code":31,"score":111,"supported":45,"title":448,"abstract":449,"tags":450},[],[28,49],[26,28,434],{"code":61,"score":188,"supported":45,"title":452,"abstract":453,"tags":454},[],[25,264],[25],[456,457],"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":459,"openalex_id":461,"authors":462,"venue":426,"cited_by_count":87,"oa_url":488,"card":489,"direction":285,"ingested_from":95},"W7222603139",[463,465,467,469,471,473,475,477,480,483,486],{"name":464,"orcid":9},"Shun He",{"name":466,"orcid":9},"Sifan Li",{"name":468,"orcid":9},"Xinjian Li",{"name":470,"orcid":9},"Qianqian Luo",{"name":472,"orcid":9},"Dong Sufen",{"name":474,"orcid":9},"Lu Lei",{"name":476,"orcid":9},"Xingjiang Chen",{"name":478,"orcid":479},"Liuti Cai","https:\u002F\u002Forcid.org\u002F0000-0002-9151-8002",{"name":481,"orcid":482},"Jian‐Yu Meng","https:\u002F\u002Forcid.org\u002F0000-0002-5796-1086",{"name":484,"orcid":485},"Ning Lu","https:\u002F\u002Forcid.org\u002F0000-0003-0214-5038",{"name":487,"orcid":9},"Hancheng Wang","https:\u002F\u002Fwww.frontiersin.org\u002Fjournals\u002Fplant-science\u002Farticles\u002F10.3389\u002Ffpls.2026.1871587\u002Fpdf",{"tldr":490,"method":491,"finding":492,"direction":92,"opportunity":493},"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"]