[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-3119":3,"related-3119":46},{"id":4,"title":5,"url":6,"summary":7,"summary_zh":8,"content":8,"source_name":9,"source_url":8,"published_at":10,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":14,"score_detail":15,"sources":23,"tags":25,"search_phrases":31,"slug":34,"view_count":35,"doi":8,"paper":36,"created_at":45},3119,"CropRowSeg: A Lightweight Model for Seedling Crop Row Image Segmentation Combining Multi-scale CNN and Transformer（融合多尺度 CNN 与 Transformer 的苗期作物行图像轻量化分割模型）","https:\u002F\u002Fnews.cau.edu.cn\u002Fkxyj\u002F9f1540e749e745f29f11355eac1c9444.htm","中国农业大学农机装备智能化设计与制造高水平创新团队翟志强副教授在国际学术期刊《Agriculture Communications》发表研究论文。中国农业大学为唯一署名单位，硕士研究生田永浩为论文第一作者、翟志强为通讯作者。研究面向农机对行作业和作物表型观测对作物行视觉感知的实际需求，提出从图像高效标注、模型架构设计到模型训练的系统解决方案：从数据标注、模型架构、损失函数三个维度构建完整的作物行图像分割技术体系；提出基于中心线引导的条带标注方法仅需人工标注冠层中心线两个端点即可自动生成分割掩码；CropRowSeg 模型采用编码器—解码器架构（编码器 CNN+ViT 双分支结构+通道-空间混合注意力机制，解码器残差连接与坐标注意力结合的渐进式解码架构）。研究可为大田作物感知提供视觉基础模型，为研发农业人工智能专用模型提供借鉴思路。",null,"中国农业大学","2026-09-18T12:39:00Z","论文",10,false,78,{"impact":16,"substance":17,"depth":18,"authority":19,"freshness":20,"relevant":21,"comment":22},18,21,17,14,8,1,"中国农业大学团队提出轻量化作物行分割模型CropRowSeg，方法体系完整、创新点明确，对农机对行作业与作物表型感知有实用价值，值得进入每日精选。",[24],{"name":9,"url":6},[26,27,28,29,30],"智慧农业","农业人工智能","农机导航","作物行识别","图像分割",[32,33],"中国农业大学 翟志强 作物行分割","CropRowSeg 苗期作物行","中国农业大学翟志强作物行分割-3119",0,{"doi":8,"openalex_id":8,"authors":37,"venue":8,"cited_by_count":35,"oa_url":8,"card":38,"direction":42,"ingested_from":44},[],{"tldr":39,"method":40,"finding":41,"direction":42,"opportunity":43},"提出轻量级作物行分割模型CropRowSeg，并设计中心线引导的条带标注方法。","多尺度CNN与ViT双分支编码器、混合注意力、渐进式解码器，中心线引导标注。","模型实现高效作物行图像分割，降低标注成本，为农机对行作业提供视觉基础。","农业遥感与作物表型","可探索该轻量模型在边缘设备上的实时部署及多作物、多生长阶段的泛化能力。","agent","2026-09-22T00:05:37.930057Z",{"total":47,"page":21,"page_size":47,"items":48},6,[49,88,137,181,228,261],{"id":50,"title":51,"url":52,"summary":53,"summary_zh":54,"content":8,"source_name":55,"source_url":52,"published_at":56,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":57,"score_detail":58,"sources":63,"tags":65,"search_phrases":68,"slug":71,"view_count":35,"doi":72,"paper":73,"created_at":87},3019,"PSPE-UNet: Projection-based Similarity Prototype Embedding UNet for Apple Leaf Disease Segmentation","https:\u002F\u002Fdoi.org\u002F10.22266\u002Fijies2026.1031.18","Apple leaf disease segmentation plays a significant role in precision agriculture by enabling the accurate identification and localization of infected regions at the pixel level.However, diverse apple leaf diseases exhibit similar symptoms, such as overlapping lesions makes it challenging to distinguish between various disease classes.In this research, a Projection-based Similarity Prototype Embedding UNet (PSPE-UNet) is proposed to segment apple leaf diseases.Employing a projection head with a similarity prototype embedding in UNet enhances feature discrimination by mapping pixel-level representations into a normalized embedding space.This ensures better separation between healthy and disease regions, even when the regions exhibit similar texture and chromatic characteristics.Three learnable prototypes corresponding to healthy, disease, and boundary regions are used.The boundary prototype act as learnable auxiliary feature prototype within the auxiliary boundary branch to compute boundary probability map during training while disease prediction is based on healthy and disease prototypes.In addition, this method enhances the boundary delineation for irregular and small lesions by refining the feature alignment.Hence, the proposed PSPE-UNet achieves a high Pixel Accuracy (PA) of 98.96%, which is compared to existing methods such as the AS-DeepLabV3+ on the Apple Tree Leaf Disease Segmentation Dataset (ATLDSD).Moreover, proposed PSPE-UNet obtains an inference time of 0.0217s per batch (8 images), corresponding to 0.0027s per image on ATLDSD dataset compared to traditional methods like UNet.","苹果叶片病害分割在精准农业中具有重要意义，能够在像素级别上准确识别和定位感染区域。然而，不同苹果叶片病害表现出相似的症状，例如病灶重叠使得区分不同病害类别具有挑战性。本研究提出了一种基于投影的相似性原型嵌入UNet（PSPE-UNet）用于苹果叶片病害分割。在UNet中采用带有相似性原型嵌入的投影头，通过将像素级表示映射到归一化嵌入空间来增强特征判别能力。这确保了健康和病害区域之间更好的分离，即使这些区域表现出相似的纹理和色彩特征。使用三个可学习原型分别对应健康、病害和边界区域。边界原型在辅助边界分支中作为可学习辅助特征原型，在训练期间计算边界概率图，而病害预测则基于健康和病害原型。此外，该方法通过细化特征对齐增强了对不规则和小病灶的边界描绘。因此，所提出的PSPE-UNet在苹果树叶病害分割数据集（ATLDSD）上达到了98.96%的高像素精度（PA），并与现有方法如AS-DeepLabV3+进行了比较。此外，所提出的PSPE-UNet在ATLDSD数据集上获得了每批次（8张图像）0.0217秒的推理时间，相当于每张图像0.0027秒，与UNet等传统方法相比具有优势。","International journal of intelligent engineering and systems","2026-09-19T00:00:00Z",70,{"impact":59,"substance":60,"depth":18,"authority":59,"freshness":61,"relevant":21,"comment":62},12,20,9,"提出基于相似度原型嵌入的UNet分割方法，在苹果叶病害数据集上取得98.96%像素精度，方法新颖、数据明确，但属细分算法研究，产业影响有限。",[64],{"name":55,"url":52},[26,27,66,30,67],"精准农业","苹果病害",[69,70],"PSPE-UNet 苹果叶病害 分割","苹果叶病害 数据集 ATLDSD","PSPE-UNet苹果叶病害分割-3019","10.22266\u002Fijies2026.1031.18",{"doi":72,"openalex_id":74,"authors":75,"venue":55,"cited_by_count":35,"oa_url":52,"card":80,"direction":84,"ingested_from":86},"W7213634285",[76,78],{"name":77,"orcid":8},"Vedamurthy Hadavanahalli Kumaraiah",{"name":79,"orcid":8},"Shrinivasacharya Purohit",{"tldr":81,"method":82,"finding":83,"direction":84,"opportunity":85},"提出PSPE-UNet，用投影相似原型嵌入分割苹果叶片病害区域。","UNet+投影头+可学习原型（健康\u002F病害\u002F边界），ATLDSD数据集。","像素精度98.96%，单图推理0.0027秒，优于AS-DeepLabV3+和UNet。","农业人工智能与决策模型","可探索原型嵌入在相似症状多病害区分及轻量化边缘部署中的泛化能力。","openalex","2026-09-20T23:30:34.933307Z",{"id":89,"title":90,"url":91,"summary":92,"summary_zh":93,"content":8,"source_name":94,"source_url":91,"published_at":95,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":96,"score_detail":97,"sources":101,"tags":103,"search_phrases":106,"slug":109,"view_count":35,"doi":110,"paper":111,"created_at":136},2316,"Vision Transformers Enable Advanced Plant Phenotyping in Controlled Environments","https:\u002F\u002Fdoi.org\u002F10.64898\u002F2026.09.04.748299","Reliable plant segmentation in high-throughput phenotyping must transfer across species and imaging conditions without repeated model tuning or extensive reannotation. We compare three segmentation strategies using images from Oak Ridge National Laboratory's Advanced Plant Phenotyping Laboratory: (i) fixed color-based thresholding, (ii) supervised U-Nets trained from scratch, and (iii) pretrained vision transformers fine-tuned for binary segmentation. Models were evaluated on a held-out test set and a generalization set that comprised unseen species. On the held-out test set, thresholding, the best U-Net, and the best vision transformer achieved mean Dice scores of 58.3, 96.6, and 97.3, respectively. On the generalization set, the corresponding Dice scores were 56.5, 86.2, and 95.7. Thresholding remained effective on some datasets but failed when plant appearance changed. Supervised U-Net training resolved within-distribution errors but failed to generalize to novel species and backgrounds. Pretrained vision transformers consistently produced high-accuracy segmentations across the evaluated species, views, soil backgrounds, and tray types. These results benchmark the practical progression from fixed rules to task-specific supervision and pretrained visual representations for controlled-environment plant phenotyping.","在高通量表型分析中，可靠的植物分割必须能够在不同物种和成像条件之间迁移，而无需重复调整模型或进行大量重新标注。我们利用橡树岭国家实验室先进植物表型实验室的图像，比较了三种分割策略：（i）基于颜色的固定阈值法，（ii）从零开始训练的有监督U-Net，以及（iii）针对二值分割进行微调的预训练视觉Transformer。模型在留出测试集和包含未见物种的泛化集上进行了评估。在留出测试集上，阈值法、最佳U-Net和最佳视觉Transformer的平均Dice分数分别为58.3、96.6和97.3。在泛化集上，相应的Dice分数分别为56.5、86.2和95.7。阈值法在某些数据集上仍然有效，但当植物外观发生变化时则失效。有监督U-Net训练解决了分布内误差，但未能泛化到新物种和背景。预训练视觉Transformer在所评估的物种、视角、土壤背景和托盘类型上始终产生高精度分割。这些结果基准了受控环境植物表型分析中从固定规则到任务特定监督再到预训练视觉表示的实际进展。","bioRxiv (Cold Spring Harbor Laboratory)","2026-09-10T00:00:00Z",79,{"impact":16,"substance":98,"depth":16,"authority":99,"freshness":20,"relevant":21,"comment":100},22,13,"预训练视觉Transformer在跨物种植物分割上显著优于U-Net与阈值法，为受控环境高通量表型提供可复用基准，方法新颖、数据扎实，值得入选。",[102],{"name":94,"url":91},[26,27,104,105,30],"高通量表型","植物表型",[107,108],"农业人工智能 高通量表型 图像分割 智慧农业","农业人工智能 高通量表型","农业人工智能高通量表型图像分割智慧农业-2316","10.64898\u002F2026.09.04.748299",{"doi":110,"openalex_id":112,"authors":113,"venue":94,"cited_by_count":35,"oa_url":8,"card":131,"direction":42,"ingested_from":86},"W7212309344",[114,116,119,122,125,128],{"name":115,"orcid":8},"Janou Milligan",{"name":117,"orcid":118},"Anand Seethepalli","https:\u002F\u002Forcid.org\u002F0000-0003-0937-9128",{"name":120,"orcid":121},"Aristeidis Tsaris","https:\u002F\u002Forcid.org\u002F0000-0002-7734-3349",{"name":123,"orcid":124},"Xiao Wang","https:\u002F\u002Forcid.org\u002F0000-0001-6545-1943",{"name":126,"orcid":127},"Larry M. York","https:\u002F\u002Forcid.org\u002F0000-0002-1995-9479",{"name":129,"orcid":130},"John Lagergren","https:\u002F\u002Forcid.org\u002F0000-0002-8092-7433",{"tldr":132,"method":133,"finding":134,"direction":42,"opportunity":135},"比较阈值法、U-Net和预训练ViT在植物分割中的跨物种泛化能力。","使用ORNL表型实验室图像，对比阈值法、U-Net和微调ViT的分割性能。","预训练ViT在未见物种上Dice达95.7，显著优于U-Net的86.2和阈值法的56.5。","可探索预训练ViT在田间复杂场景的跨物种泛化，并降低微调数据需求。","2026-09-13T23:30:21.474320Z",{"id":138,"title":139,"url":140,"summary":141,"summary_zh":142,"content":8,"source_name":143,"source_url":140,"published_at":144,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":145,"score_detail":146,"sources":150,"tags":152,"search_phrases":153,"slug":156,"view_count":35,"doi":157,"paper":158,"created_at":180},1631,"A dynamic multi-scale feature fusion and hierarchical attention network for leaf segmentation","https:\u002F\u002Fdoi.org\u002F10.1007\u002Fs10791-026-10494-2","Abstract Accurate leaf instance segmentation is fundamental to quantifying morphological and structural traits in image-based plant phenotyping. However, substantial variations in leaf scale, shape, and orientation, together with dense overlap and occlusion, often lead to ambiguous instance boundaries. In addition, repeated downsampling and feature reconstruction can progressively erode fine structural details, hindering contour preservation and the separation of adjacent leaves. To address these interrelated challenges, we propose DMSHA-Net, a dynamic multi-scale feature fusion and hierarchical attention network for leaf instance segmentation. Its direction-aware Multi-Scale Feature Aggregation (MSFA) encoder captures complementary horizontal and vertical contextual information across multiple receptive-field scales, thereby improving the representation of diverse leaf morphologies. The Dense Feature Aggregation (DFA) decoder selectively integrates deep semantic and shallow structural features through stage-specific attention mechanisms, where self-attention at coarse resolutions models long-range dependencies, whereas lightweight channel recalibration at high resolutions refines local structures and boundaries. The Learnable Feature Fusion (LFF) module subsequently integrates multi-level semantic and boundary features using normalized learned weights. DMSHA-Net achieves Best Dice (BD) scores of 93.17%, 85.12%, and 90.91% on KOMATSUNA, MSU-PID, and the CVPPP-A1 subset, respectively, with corresponding foreground–background Dice (FBD) scores of 98.19%, 91.02%, and 98.24%. These results indicate that DMSHA-Net achieves competitive segmentation performance across the three datasets, particularly in scenarios involving pronounced scale variation and dense leaf overlap.","准确的叶片实例分割是基于图像植物表型分析中量化形态与结构特征的基础。然而，叶片在尺度、形状和方向上的显著差异，加之密集重叠和遮挡，常导致实例边界模糊不清。此外，重复的下采样和特征重建会逐步侵蚀细微的结构细节，妨碍轮廓保持及相邻叶片的分离。为应对这些相互关联的挑战，我们提出了DMSHA-Net，一种用于叶片实例分割的动态多尺度特征融合与层次注意力网络。其方向感知的多尺度特征聚合（MSFA）编码器在多个感受野尺度上捕获互补的水平和垂直上下文信息，从而增强对不同叶片形态的表征能力。密集特征聚合（DFA）解码器通过阶段特定的注意力机制选择性地整合深层语义特征与浅层结构特征，其中在粗分辨率下的自注意力建模长距离依赖关系，而在高分辨率下的轻量级通道重校准则细化局部结构和边界。可学习特征融合（LFF）模块随后利用归一化的学习权重整合多层次语义与边界特征。DMSHA-Net在KOMATSUNA、MSU-PID和CVPPP-A1子集上分别取得了93.17%、85.12%和90.91%的最佳Dice（BD）分数，相应的前景-背景Dice（FBD）分数分别为98.19%、91.02%和98.24%。这些结果表明，DMSHA-Net在三个数据集上均实现了具有竞争力的分割性能，尤其在涉及显著尺度变化和密集叶片重叠的场景中表现突出。","Discover Computing","2026-09-03T00:00:00Z",74,{"impact":147,"substance":98,"depth":16,"authority":59,"freshness":148,"relevant":21,"comment":149},15,7,"提出DMSHA-Net网络，在多个植物叶片分割数据集上取得领先性能，对农业表型分析有实质推进。",[151],{"name":143,"url":140},[26,27,105,30],[154,155],"农业人工智能 图像分割 智慧农业 植物表型","农业人工智能 图像分割","农业人工智能图像分割智慧农业植物表型-1631","10.1007\u002Fs10791-026-10494-2",{"doi":157,"openalex_id":159,"authors":160,"venue":143,"cited_by_count":35,"oa_url":140,"card":175,"direction":42,"ingested_from":86},"W7207602965",[161,164,166,168,171,173],{"name":162,"orcid":163},"Fei Liu","https:\u002F\u002Forcid.org\u002F0000-0002-6148-3697",{"name":165,"orcid":8},"Yingjie Fan",{"name":167,"orcid":8},"Mingtao Zhou",{"name":169,"orcid":170},"Huabing Liu","https:\u002F\u002Forcid.org\u002F0000-0003-1312-3926",{"name":172,"orcid":8},"Shuya Chen",{"name":174,"orcid":8},"Bin Wen",{"tldr":176,"method":177,"finding":178,"direction":42,"opportunity":179},"提出动态多尺度特征融合与层级注意力网络，用于叶片实例分割，提升精度。","使用方向感知多尺度特征聚合编码器、密集特征聚合解码器及可学习特征融合模块。","在三个数据集上取得最佳Dice分数，尤其在尺度变化和密集重叠场景表现优异。","可探索将DMSHA-Net应用于更多作物类型或复杂田间环境，或结合轻量化设计用于实时表型分析。","2026-09-04T23:30:26.243015Z",{"id":182,"title":183,"url":184,"summary":185,"summary_zh":186,"content":8,"source_name":187,"source_url":184,"published_at":188,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":145,"score_detail":189,"sources":191,"tags":193,"search_phrases":196,"slug":199,"view_count":35,"doi":200,"paper":201,"created_at":227},1510,"From frame-wise detection to instance-level tracking: robust crop row perception for autonomous paddy field navigation","https:\u002F\u002Fdoi.org\u002F10.1016\u002Fj.compag.2026.112238","From frame-wise detection to instance-level tracking: robust crop row perception for autonomous paddy field navigation。Computers and Electronics in Agriculture","从逐帧检测到实例级跟踪：面向自主稻田导航的稳健作物行感知。《农业计算机与电子》","Computers and Electronics in Agriculture","2026-09-02T00:00:00Z",{"impact":147,"substance":60,"depth":16,"authority":19,"freshness":148,"relevant":21,"comment":190},"论文提出从帧级检测到实例级跟踪的作物行感知方法，提升水稻田间自主导航鲁棒性，对农业机器人领域有实质贡献。",[192],{"name":187,"url":184},[26,27,194,195,29],"水稻","田间导航",[197,198],"农业人工智能 作物行识别 智慧农业 田间导航","农业人工智能 作物行识别","农业人工智能作物行识别智慧农业田间导航-1510","10.1016\u002Fj.compag.2026.112238",{"doi":200,"openalex_id":202,"authors":203,"venue":187,"cited_by_count":35,"oa_url":8,"card":222,"direction":84,"ingested_from":86},"W7206160448",[204,207,209,212,215,217,219],{"name":205,"orcid":206},"Dongfang Li","https:\u002F\u002Forcid.org\u002F0009-0008-1790-9022",{"name":208,"orcid":8},"Feiyang Zhang",{"name":210,"orcid":211},"Xingshuo Li","https:\u002F\u002Forcid.org\u002F0000-0002-1262-406X",{"name":213,"orcid":214},"Yongkai Ye","https:\u002F\u002Forcid.org\u002F0000-0002-2945-2065",{"name":216,"orcid":8},"Jun Wang",{"name":218,"orcid":8},"Yejun Zhu",{"name":220,"orcid":221},"Maohua Xiao","https:\u002F\u002Forcid.org\u002F0000-0001-5213-1035",{"tldr":223,"method":224,"finding":225,"direction":84,"opportunity":226},"提出从逐帧检测到实例级跟踪的作物行感知方法，用于水稻田自主导航。","结合检测与跟踪技术，实现作物行实例级感知。","该方法能提升水稻田导航的鲁棒性。","可探索在复杂农田环境下，结合多传感器与深度学习的作物行跟踪与导航方法。","2026-09-03T23:30:02.171311Z",{"id":229,"title":230,"url":231,"summary":232,"summary_zh":8,"content":8,"source_name":233,"source_url":8,"published_at":234,"category":11,"cover_url":8,"hotness":12,"is_selected":235,"score":236,"score_detail":237,"sources":242,"tags":244,"search_phrases":249,"slug":252,"view_count":35,"doi":8,"paper":253,"created_at":260},3125,"Full-Season Agentic Farm System FAIRY: Event-Driven Multi-Agent Orchestration for Soybean Production（FAIRY 全季节智能体农场系统：大豆生产的事件驱动多智能体编排）","https:\u002F\u002Faiagentstore.ai\u002Fai-agent-news\u002Ftopic\u002Fagriculture-food\u002F2026-08-11","哈尔滨工业大学研究人员发布并部署全栈、事件驱动的智能体引擎 FAIRY 于中国运行中的大豆研究农场。FAIRY 集成传感器、无人机、卫星植被产品、机械 API、作物过程模型与多智能体编排层，执行起垄→播种→灌溉→病虫害防治→收获→干燥→存储工作流，并在 64 垄研究场上跨 100 个全季节场景评估 9 个智能体控制器。这是智能体系统能够在大农业时间尺度和延迟结果下进行推理的最清晰演示之一，将农业中的智能体工作从实验室演示推进到全过程评估。同期 arXiv 推出 HarvestBench 基准将 LLM 驱动智能体置于农场网格世界（拖拉机面临动物选择绕行或碾压），结果显示模型差异巨大、对道德简报高度敏感、避免意愿具有可衡量的价格弹性。","Harbin Institute of Technology \u002F arXiv","2026-09-18T00:00:00Z",true,89,{"impact":238,"substance":239,"depth":240,"authority":19,"freshness":61,"relevant":21,"comment":241},24,23,19,"哈工大在真实大豆农场部署全季节事件驱动多智能体系统并配套 HarvestBench 基准，是农业智能体从演示走向全过程评估的标志性进展，专业深度与信息增量俱佳。",[243],{"name":233,"url":231},[26,245,27,246,247,248],"无人农场","农业遥感","多智能体","大豆生产",[250,251],"哈工大 FAIRY 大豆","HarvestBench 智能体 农场","哈工大FAIRY大豆-3125",{"doi":8,"openalex_id":8,"authors":254,"venue":8,"cited_by_count":35,"oa_url":8,"card":255,"direction":84,"ingested_from":44},[],{"tldr":256,"method":257,"finding":258,"direction":84,"opportunity":259},"部署全季节事件驱动多智能体系统FAIRY，在大豆农场完成从起垄到存储的全流程评估。","集成传感器、无人机、卫星、作物模型与机械API，用9个智能体控制器在64垄100","智能体系统能在大农业时间尺度下推理，模型差异大且对道德简报敏感。","可研究多智能体在长周期、延迟反馈农业任务中的鲁棒性与伦理约束机制。","2026-09-22T00:05:38.611717Z",{"id":262,"title":263,"url":264,"summary":265,"summary_zh":8,"content":8,"source_name":266,"source_url":8,"published_at":234,"category":11,"cover_url":8,"hotness":12,"is_selected":13,"score":14,"score_detail":267,"sources":269,"tags":271,"search_phrases":275,"slug":278,"view_count":35,"doi":8,"paper":279,"created_at":287},3120,"Autonomous Agricultural Machinery for Smart Agriculture: An Integrated Framework of Observation, Heterogeneity, and Intelligent Infrastructure（智慧农业自主农机：观测、异质性与智能基础设施一体化框架）","https:\u002F\u002Fwww.jstage.jst.go.jp\u002Farticle\u002Frdj\u002F5\u002F0\u002F5_576\u002F_article\u002F-char\u002Fen","Hongjin Li & Chunjiang Gao 在 Journal of Agricultural Information 5: 576-596 发表综述文章（J-STAGE 平台 2026-09-18 上线）：针对农业 4.0 时代自主农机核心问题——如何在复杂动态田间条件下将异构环境信息转化为可靠自适应可扩展的机器自主性，系统综述自主农机的技术基础、应用领域、性能优势与采用约束，并构建围绕\"观测、异质性、基础设施\"的一体化框架。研究不再按机器类型分类技术，而是考察多模态感知、AI 决策、自主导航控制、精准执行、多机协同之间的耦合关系，特别关注环境不确定性、实时决策、互操作性和系统级可扩展性的挑战。识别从孤立任务自动化向数据驱动自适应网络化农业自主性的渐进过渡，强调个体组件改进未必带来系统级性能提升，除非具备兼容的计算、通信、机械和制度基础设施。","J-STAGE \u002F Research Disclosure Journal",{"impact":16,"substance":17,"depth":18,"authority":99,"freshness":61,"relevant":21,"comment":268},"核心期刊综述，提出观测-异质性-基础设施一体化框架，对智慧农业自主农机研究有参考价值，但属学术综述、产业影响有限。",[270],{"name":266,"url":264},[26,27,272,273,274],"多模态感知","自主农机","农业4.0",[276,277],"自主农机 智慧农业 一体化框架","Hongjin Li Chunjiang Gao 自主农机","自主农机智慧农业一体化框架-3120",{"doi":8,"openalex_id":8,"authors":280,"venue":8,"cited_by_count":35,"oa_url":8,"card":281,"direction":285,"ingested_from":44},[],{"tldr":282,"method":283,"finding":284,"direction":285,"opportunity":286},"综述自主农机技术，提出观测、异质性、基础设施一体化框架。","系统综述多模态感知、AI决策、导航控制与多机协同的耦合关系。","组件改进未必提升系统性能，需兼容的计算、通信、机械与制度基础设施。","智慧农业 \u002F 农业物联网","可研究异构农机互操作协议与边缘计算架构，填补系统级可扩展性验证空白。","2026-09-22T00:05:38.042546Z"]