[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-472":3,"related-472":51},{"id":4,"title":5,"url":6,"summary":7,"summary_zh":8,"content":9,"source_name":10,"source_url":8,"published_at":11,"category":12,"cover_url":8,"hotness":13,"is_selected":14,"score":15,"score_detail":16,"sources":23,"tags":25,"primary_topic":8,"secondary_topics":30,"classification":31,"search_phrases":45,"slug":48,"view_count":49,"doi":8,"paper":8,"created_at":50},472,"Multi-Agent Vision-Language Debate Framework for Zero-Shot Crop Disease Diagnosis (VIDA+PANDA)","https:\u002F\u002Fwww.ebiotrade.com\u002Fnewsf\u002F2026-8\u002F20260812000305146.htm","本文提出VIDA+PANDA多智能体视觉语言框架用于零样本作物病害诊断,包含两个阶段:VIDA多VLM智能体独立分析作物叶片图像建立基线性能;同行锚定命名审议架构(PANDA)在选定的智能体群体中引入结构化多轮辩论,智能体交换推理、批评同行预测、通过基于证据的讨论修订决策,反谄媚机制阻止无支持的共识转变。在CDDM基准上使用来自四个独立提供商的七个异构VLM,在完全零样本设置下,非参与方GPT-5作为独立评判者评估。框架引入三个语义度量:语义标签相似度(SLS)、推理特异性(RS)、智能体间推理收敛度(IRC)。实验结果表明协作多智能体审议提高了个体诊断性能和语义对齐,较弱智能体增益最大。原文框架发表于Frontiers in Plant Science。",null,"《Frontiers in Plant Science》：A multi-agent vision-language debate framework for zero-shot crop disease diagnosis\n\n准确的作物病害诊断对于农业生产力和粮食安全至关重要，然而现有的深度学习系统通常难以在视觉相似的疾病和变化的环境条件下泛化。最近的视觉语言模型（VLMs）已展现出有前景的零样本推理能力；然而，大多数农业诊断系统仍依赖于孤立的单模型预测，缺乏协作推理或共识机制。在本工作中，研究人员提出了VIDA+PANDA，一种用于零样本作物病害诊断的多智能体视觉语言框架。该框架包含两个阶段：VIDA，其中多个VLM智能体独立分析作物叶片图像以建立基线性能；以及同行锚定命名审议架构（PANDA），该架构在选定的高性能和架构多样的智能体群体中引入结构化的多轮辩论。在审议过程中，智能体交换推理、批评同行预测，并通过基于证据的讨论修订决策，同时一个反谄媚机制阻止无支持的共识转变。最终预测通过性能加权共识投票生成。实验在CDDM基准上进行，使用来自四个独立提供商的七个异构VLM，包括两个开源模型，在完全零样本设置下进行。一个非参与方的GPT-5模型作为独立评判者评估最终诊断预测。除了传统准确性，该框架引入了三个语义度量：语义标签相似度（SLS），衡量预测的作物-病害对与真实标签的语义接近程度，并捕获精确匹配评估忽略的部分正确性；推理特异性（RS），衡量智能体的解释具体引用视觉证据（如病斑颜色、形状、纹理或边缘）的程度；以及智能体间推理收敛度（IRC），衡量智能体依赖相似视觉证据的程度，独立于标签正确性捕获认知对齐。实验结果表明，协作多智能体审议提高了个体诊断性能和语义对齐，其中在较弱的参与智能体中观察到最大的增益。研究结果还揭示了基于VLM的农业诊断中预测准确性、说服性影响和共识形成之间的重要关系。\n\n**论文解读：零样本作物病害诊断的多智能体视觉语言辩论框架**\n\n**研究背景与问题** \n\n农业是全球粮食安全的基石，但面临气候变化、劳动力短缺和病原体演变等挑战，作物病害每年造成约30%的产量损失。早期准确诊断至关重要，但传统方法（专家目视或实验室分析）耗时费力，难以大规模部署。深度学习在农业病害检测中取得进展，但现有系统存在三大局限：一是依赖大量标注数据，泛化能力差；二是面对视觉相似疾病和环境变化时性能下降；三是多数系统采用单模型推理，缺乏协作推理或共识机制。视觉语言模型（VLMs）虽具备零样本推理能力，但单个VLM存在幻觉、推理不一致和过度自信等问题。因此，研究人员提出多智能体辩论框架，探索结构化协作推理能否提升零样本诊断的可靠性和语义对齐。\n\n**研究内容与结论** \n\n研究人员提出了VIDA+PANDA框架，包含两个阶段：VIDA阶段（独立诊断）和PANDA阶段（结构化多轮命名辩论）。在CDDM基准（16种作物、54种病害类别，1,219张图像，分层抽样80%患病20%健康）上，使用来自四个独立提供商的七个异构VLM（包括GPT-4.1、Grok-4.20-Fast、GPT-4.1-mini、Claude Sonnet 4.5、Qwen-VL、Claude Haiku 4.5、Gemma-3n-E4B，其中两个开源模型）进行完全零样本评估。非参与方GPT-5作为独立评判者评估诊断预测。框架引入三个语义度量：语义标签相似度（SLS）、推理特异性（RS）和智能体间推理收敛度（IRC）。研究发现：协作辩论显著提升弱智能体性能（如开源Qwen-VL综合准确率从0.183提升至0.304），但加权集成共识未改进，受限于最强锚点（GPT-4.1）；说服性影响与准确性解耦，最固执的智能体（Grok-4.20-Fast）反而最具说服力，存在“说服性不一致”风险；智能体间推理收敛度（IRC）与诊断正确性正相关，是独立于标签的正确性信号。论文发表在《Frontiers in Plant Science》。\n\n**关键方法** \n\n研究使用CDDM基准（来源：Liu等，2024b），包含1,219张图像，覆盖53个病害类别（排除番茄黄化曲叶病毒类别）。七个VLM智能体在VIDA阶段独立分析图像，输出作物、病害、推理和置信度，经三阶段解析器对齐。基于VIDA综合准确率，选择三个高性能且架构多样的智能体（GPT-4.1、Qwen-VL、Grok-4.20-Fast）进入PANDA阶段。PANDA进行两轮命名辩论：Round 2中智能体收到同行预测和推理摘要，要求引用新视觉证据才能改变预测（反谄媚机制）；Round 3中智能体基于完整历史进行最终决策。最终预测通过性能加权共识（softmax权重，温度T=2.0）生成。非参与方GPT-5评判推理质量和合理性，并计算三个语义度量（SLS、RS、IRC）。所有实验使用固定种子42，数据集名称从提示词中隐藏。\n\n**研究结果** \n\n**5.1 VIDA：独立诊断性能** \n\n通过分析七个智能体在1,219张图像上的独立诊断结果，研究人员发现GPT-4.1综合准确率最高（0.366），其次是Grok-4.20-Fast（0.326）和GPT-4.1-mini（0.303）。开源Qwen-VL（0.184）居中。所有智能体置信度校准均较差（Pearson r在0.09至0.39之间），表明置信度不能可靠反映正确性。 \n\n**5.1.1 每作物难度** \n\n通过Kruskal-Wallis检验（H=830.12，p≈3.1×10-167），作物间准确率差异显著：视觉上独特的作物（小麦0.952、覆盆子0.895、南瓜0.880）诊断可靠，而混淆疾病多的作物（水稻0.127、樱桃0.068、桃0.103）即使最强智能体也表现差。 \n\n**5.1.2 每病害难度** \n\n通过汇总所有智能体的综合准确率，最难的病害（如Target Spot、Grey Spot、Leaf Smut）作物准确率高但病害准确率极低（接近0），表明瓶颈在于细粒度疾病区分而非作物识别。健康类准确率（0.506）远高于患病类（0.156），提示二元疾病分诊更易处理。 \n\n**5.1.3 样本量鲁棒性** \n\n通过比较全量1,219张图像与分层490张子样本的准确率，每个智能体变化不超过0.014，在Wilson置信区间内，表明结论稳定。 \n\n**5.2 PANDA：辩论动态与共识** \n\n**5.2.1 共识锚定不变** \n\n通过McNemar检验，辩论前后加权共识综合准确率无显著变化（0.359→0.360，p=1.000），作物和疾病准确率也无显著变化。共识被最强锚点（GPT-4.1）绑定，不因辩论改善。 \n\n**5.2.2 个体智能体变化显著** \n\n通过Wilcoxon符号秩检验，开源Qwen-VL综合准确率从0.183提升至0.304（p\u003C0.0001），修订575张图像的疾病判断，净改善122例。GPT-4.1微升（+0.010，p=0.134），Grok几乎不变（-0.003，p=0.248）。辩论收益集中在弱智能体。 \n\n**5.2.3 说服不对称且与准确性解耦** \n\n通过分析智能体被引用的次数，GPT-4.1说服Qwen-VL 882次，Grok说服Qwen-VL 324次，而Qwen几乎不说服别人。Grok是最固执的（仅修订1.3%），但极具说服力，表明说服性影响不等于正确性。 \n\n**5.3 智能体配置消融** \n\n**5.3.1 辩论前后准确率** \n\n通过比较四种三智能体面板（PANDA-1至4）在490张子集上的表现，包含GPT-4.1的面板（PANDA-1）综合准确率提升最大（+0.017），而不含GPT-4.1的面板（PANDA-3）无变化（0.286）。McNemar检验均未达显著，但PANDA-2作物准确率显著下降（-0.031，p=0.005），由弱智能体（Claude-Sonnet）说服强锚点（GPT-4.1）导致。 \n\n**5.3.2 每智能体轨迹与锚点收敛** \n\n通过分析每轮准确率，弱智能体（GPT-4.1-mini、Claude-Sonnet）显著提升（p\u003C0.001），但收敛于锚点准确率，个体增益在集成层面抵消。 \n\n**5.3.3 跨面板的同行影响** \n\n通过对比PANDA-4（含GPT-4.1）和PANDA-3（不含），含锚点时共识保持0.369，不含时共识停滞0.286，尽管弱智能体个体提升。辩论仅在有高准确率锚点时才能改善集成。 \n\n**5.4 推理收敛的语义评估** \n\n通过分析PANDA-4配置下的三个语义度量，SLS从5.18提升至5.68（Δ+0.49），RS从7.79下降至7.46（Δ-0.33，因后期响应包含社交语言），IRC从5.80提升至7.25（Δ+1.45）。共识正确的图像IRC更高（≈8.0 vs 7.0），表明推理收敛是可靠诊断的信号。\n\n**讨论与结论** \n\n讨论指出，辩论帮助个体而非集成，共识受最强锚点绑定；说服性不一致是主要风险，即最固执的智能体可能最不准确但最具说服力；开源模型可行且成本可控，但成本仍为实际部署的约束；细粒度错误集中在少数疾病类别，可通过少量样本提示改进；局限性包括使用单叶图像、未测试田间条件，语义评分需人类专家验证。结论部分翻译如下： \n\n研究人员提出了VIDA+PANDA，一种用于作物病害诊断的零样本、异构多智能体视觉语言框架，并在1,219张图像、53个类别上使用非参与方GPT-5评判者、隐藏数据集名称的提示词进行评估。两个发现突出：首先，结构化同行辩论未改善性能加权集成共识，后者仍受最强锚点限制，但显著提升最弱个体智能体（一个开源模型），表明辩论提高下限而非上限；其次，同行影响强烈不对称且与准确性解耦：最具说服力的智能体未必最正确或最愿意更新。这种说服性不一致是多智能体诊断辩论的主要风险，必须在实际部署前加以控制。此外，研究人员引入了三个语义度量（SLS、RS、IRC），并展示了智能体间推理收敛度追踪诊断正确性，提供了独立于标签一致性的可靠性信号。未来工作针对田间图像、少数未解决细粒度类别的少量样本适应、可靠性加权影响以抑制说服性不一致，以及与分割管道集成用于温室验证。","Frontiers in Plant Science 2026-08-12","2026-08-12T00:00:00Z","论文",10,false,84,{"impact":17,"substance":18,"depth":19,"authority":20,"freshness":13,"relevant":21,"comment":22},18,22,20,14,1,"前沿多智能体框架提升零样本作物病害诊断，方法新颖、数据详实，对智慧农业有重要参考价值。",[24],{"name":10,"url":6},[26,27,28,29],"作物病害","多智能体","视觉语言模型","零样本",[],{"status":32,"locked":14,"version":33,"method":34,"evidence":35},"unclassified","2026-10-22.2","rules",{"candidates":36,"agricultural_evidence":43,"reason":44},[37],{"code":38,"score":39,"supported":14,"title":40,"abstract":41,"tags":42},"I2",2,[],[],[27,28],true,"缺少足够的主题证据",[46,47],"视觉语言模型 作物病害 多智能体 零样本","视觉语言模型 作物病害","视觉语言模型作物病害多智能体零样本-472",4,"2026-08-12T23:57:24.396861Z",{"total":52,"page":21,"page_size":52,"items":53},6,[54,124,203,265,305,364],{"id":55,"title":56,"url":57,"summary":58,"summary_zh":59,"content":8,"source_name":60,"source_url":57,"published_at":61,"category":12,"cover_url":8,"hotness":13,"is_selected":14,"score":62,"score_detail":63,"sources":67,"tags":69,"primary_topic":74,"secondary_topics":79,"classification":80,"search_phrases":101,"slug":104,"view_count":105,"doi":106,"paper":107,"created_at":123},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。","Zenodo (CERN European Organization for Nuclear Research)","2026-10-21T00:00:00Z",75,{"impact":64,"substance":18,"depth":17,"authority":65,"freshness":13,"relevant":21,"comment":66},12,13,"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。",[68],{"name":60,"url":57},[70,71,72,26,73],"智慧农业","农业人工智能","计算机视觉","气候变化",{"code":38,"slug":75,"name":76,"path":77,"parent_code":78},"agricultural-ai","农业AI与大模型","\u002Ftopic\u002Fdigital-agriculture\u002Fagricultural-ai","I",[],{"status":81,"locked":14,"version":33,"method":34,"evidence":82},"classified",{"candidates":83,"agricultural_evidence":43},[84,91,96],{"code":38,"score":85,"supported":43,"title":86,"abstract":89,"tags":90},24,[87,88],"artificial intelligence","computer vision",[72,88],[71,72],{"code":92,"score":21,"supported":14,"title":93,"abstract":94,"tags":95},"H7",[],[],[73],{"code":97,"score":21,"supported":14,"title":98,"abstract":99,"tags":100},"I1",[],[],[70],[102,103],"NASA POWER 哈拉雷 气温","Wasserstein 分布鲁棒 病害分类","NASAPOWER哈拉雷气温-5584",0,"10.5281\u002Fzenodo.23544150",{"doi":106,"openalex_id":108,"authors":109,"venue":60,"cited_by_count":105,"oa_url":57,"card":116,"direction":120,"ingested_from":122},"W7223087892",[110,112,114],{"name":111,"orcid":8},"Kossi Afeli",{"name":113,"orcid":8},"Akouvi Amegashie",{"name":115,"orcid":8},"Komlan Sélom Dossou",{"tldr":117,"method":118,"finding":119,"direction":120,"opportunity":121},"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。","农业人工智能与决策模型","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。","openalex","2026-10-22T11:21:00.036937Z",{"id":125,"title":126,"url":127,"summary":128,"summary_zh":129,"content":8,"source_name":130,"source_url":127,"published_at":131,"category":12,"cover_url":8,"hotness":13,"is_selected":14,"score":132,"score_detail":133,"sources":139,"tags":141,"primary_topic":144,"secondary_topics":145,"classification":146,"search_phrases":164,"slug":167,"view_count":21,"doi":168,"paper":169,"created_at":202},5147,"FDUNet-LRSA: efficient frequency-adaptive segmentation for automated crop disease inspection","https:\u002F\u002Fdoi.org\u002F10.3389\u002Ffpls.2026.1963106","The transition toward industrialized and automated agriculture requires reliable machine vision for high-throughput crop health monitoring and field decision support. Automated sweetpotato disease inspection, however, remains challenging because field images exhibit substantial variations in leaf scale, appearance, illumination, occlusion, and background complexity. Conventional U-shaped networks rely primarily on fixed convolutional kernels and local feature aggregation, whereas Transformer-based alternatives often introduce considerable computational overhead. This article proposes FDUNet-LRSA, a frequency-adaptive segmentation network that integrates frequency-domain dynamic convolution, low-resolution self-attention, and boundary-aware refinement. Frequency-domain dynamic convolution reconstructs input-dependent spatial kernels to improve feature adaptation under changing imaging conditions. Low-resolution self-attention captures long-range contextual dependencies in a fixed compact feature space, while boundary-aware refinement preserves irregular leaf contours during mask reconstruction. Under the reported evaluation protocol, FDUNet-LRSA achieved a mean test foreground IoU of 0.94 and a Dice coefficient of 0.97, with 3.72 GFLOPs. Additional experiments indicate that task-oriented moderate augmentation improves generalization compared with stronger augmentation. The proposed network provides an efficient visual perception approach for automated crop inspection systems.","向工业化与自动化农业转型，需要可靠的机器视觉技术，以支持高通量作物健康监测与田间决策。然而，甘薯病害自动检测仍具挑战性，因为田间图像在叶片尺度、外观、光照、遮挡和背景复杂度方面存在显著变化。传统U形网络主要依赖固定卷积核和局部特征聚合，而基于Transformer的替代方案往往带来较大的计算开销。本文提出FDUNet-LRSA，一种频率自适应分割网络，融合了频域动态卷积、低分辨率自注意力与边界感知细化。频域动态卷积重建依赖输入的空间卷积核，以提升成像条件变化下的特征适应性。低分辨率自注意力在固定的紧凑特征空间中捕获长程上下文依赖，而边界感知细化在掩码重建过程中保留不规则叶片轮廓。在所报告的评估协议下，FDUNet-LRSA取得了0.94的平均测试前景IoU和0.97的Dice系数，计算量为3.72 GFLOPs。额外实验表明，与更强的增强相比，面向任务的中等增强可提升泛化能力。所提出的网络为自动化作物检测系统提供了一种高效的视觉感知方法。","Frontiers in Plant Science","2026-10-16T00:00:00Z",76,{"impact":134,"substance":135,"depth":136,"authority":65,"freshness":137,"relevant":21,"comment":138},16,21,17,9,"提出频域自适应分割网络，在甘薯病害检测上取得高精度且计算量低，方法有新意、结论可靠，属细分领域实质进展，值得进入每日精选。",[140],{"name":130,"url":127},[70,71,26,142,143],"甘薯","图像分割",{"code":38,"slug":75,"name":76,"path":77,"parent_code":78},[],{"status":81,"locked":14,"version":33,"method":34,"evidence":147},{"candidates":148,"agricultural_evidence":43},[149,155,160],{"code":38,"score":150,"supported":43,"title":151,"abstract":152,"tags":154},5,[],[153],"decision support",[71,143],{"code":156,"score":49,"supported":43,"title":157,"abstract":158,"tags":159},"A5",[],[142],[142],{"code":97,"score":21,"supported":14,"title":161,"abstract":162,"tags":163},[],[],[70],[165,166],"FDUNet-LRSA 甘薯病害","甘薯病害 图像分割","FDUNet-LRSA甘薯病害-5147","10.3389\u002Ffpls.2026.1963106",{"doi":168,"openalex_id":170,"authors":171,"venue":130,"cited_by_count":105,"oa_url":194,"card":195,"direction":201,"ingested_from":122},"W7222502318",[172,175,178,181,183,186,188,190,192],{"name":173,"orcid":174},"Ru Han","https:\u002F\u002Forcid.org\u002F0009-0000-5128-1314",{"name":176,"orcid":177},"Lei Shu","https:\u002F\u002Forcid.org\u002F0000-0002-6700-9347",{"name":179,"orcid":180},"Grzegorz Cielniak","https:\u002F\u002Forcid.org\u002F0000-0002-6299-8465",{"name":182,"orcid":8},"Yan Zhu",{"name":184,"orcid":185},"Jucheng Yang","https:\u002F\u002Forcid.org\u002F0000-0002-5046-2663",{"name":187,"orcid":8},"Fan Yang",{"name":189,"orcid":8},"Weihan Song",{"name":191,"orcid":8},"Qian Zhang",{"name":193,"orcid":8},"Kai Huang","https:\u002F\u002Fwww.frontiersin.org\u002Fjournals\u002Fplant-science\u002Farticles\u002F10.3389\u002Ffpls.2026.1963106\u002Fpdf",{"tldr":196,"method":197,"finding":198,"direction":199,"opportunity":200},"提出频域自适应分割网络FDUNet-LRSA，用于甘薯病害自动巡检。","频域动态卷积、低分辨率自注意力与边界感知细化，结合适度数据增强。","测试前景IoU达0.94、Dice达0.97，仅3.72 GFLOPs，轻量高效。","农业遥感与作物表型","可探索频域动态卷积在多种作物病害与复杂田间光照下的跨域泛化与边缘部署。","智慧农业 \u002F 农业物联网","2026-10-17T23:30:08.487857Z",{"id":204,"title":205,"url":206,"summary":207,"summary_zh":8,"content":8,"source_name":208,"source_url":8,"published_at":209,"category":12,"cover_url":8,"hotness":13,"is_selected":14,"score":210,"score_detail":211,"sources":213,"tags":215,"primary_topic":219,"secondary_topics":225,"classification":231,"search_phrases":252,"slug":255,"view_count":39,"doi":8,"paper":256,"created_at":264},4648,"[预印本]面向农业机器人在多变天气与地形下基于视觉-语言模型的上下文自适应农药喷洒(Context-Aware Adaptive Pesticide Spraying for Agricultural Robots under Changing Weather and Terrain Using Vision-Language Models)","https:\u002F\u002Farxiv.org\u002Fabs\u002F2610.08807","arXiv论文(2610.08807,2026年10月6日提交)提出基于视觉-语言模型(VLM)的上下文自适应喷洒框架,通过空间推理整合作物、农药、气象数据驱动喷洒决策,并结合MPPI轨迹跟踪控制。作物行检测精度至少提升30%,机器人可在两种环境中灵活调节喷洒量与行驶速度并减少农药漂移。[预印本未经同行评议,尚未实地部署验证]。","arXiv 2026-10-06 [预印本]","2026-10-06T00:00:00Z",72,{"impact":17,"substance":19,"depth":17,"authority":13,"freshness":52,"relevant":21,"comment":212},"预印本提出VLM驱动的自适应喷洒框架，方法新颖且减药减排指向明确，但未经同行评议与实地验证，宜作前沿动态收录而非重点精选。",[214],{"name":208,"url":206},[70,216,217,28,218],"农业机器人","精准施药","农药减量",{"code":220,"slug":221,"name":222,"path":223,"parent_code":224},"F4","crop-protection","农药与绿色防控","\u002Ftopic\u002Finputs-equipment\u002Fcrop-protection","F",[226],{"code":227,"slug":228,"name":229,"path":230,"parent_code":224},"F6","agricultural-robots","机器人与无人农场","\u002Ftopic\u002Finputs-equipment\u002Fagricultural-robots",{"status":81,"locked":14,"version":33,"method":34,"evidence":232},{"candidates":233,"agricultural_evidence":43},[234,240,244,248],{"code":220,"score":135,"supported":43,"title":235,"abstract":238,"tags":239},[236,237],"农药","pesticide",[236],[218,217],{"code":227,"score":137,"supported":43,"title":241,"abstract":242,"tags":243},[216],[],[216],{"code":97,"score":21,"supported":14,"title":245,"abstract":246,"tags":247},[],[],[70],{"code":38,"score":21,"supported":14,"title":249,"abstract":250,"tags":251},[],[],[28],[253,254],"农业机器人 视觉语言模型 农药喷洒","MPPI 轨迹跟踪 作物行检测","农业机器人视觉语言模型农药喷洒-4648",{"doi":8,"openalex_id":8,"authors":257,"venue":8,"cited_by_count":105,"oa_url":8,"card":258,"direction":201,"ingested_from":263},[],{"tldr":259,"method":260,"finding":261,"direction":201,"opportunity":262},"提出基于视觉-语言模型的上下文自适应农药喷洒框架，结合MPPI控制实现精准喷洒。","视觉-语言模型空间推理，融合作物、农药与气象数据，MPPI轨迹跟踪控制。","作物行检测精度至少提升30%，可灵活调节喷洒量与速度并减少农药漂移。","预印本未实地验证，可探索真实多变天气地形下的VLM鲁棒性与多模态数据融合优化。","agent","2026-10-11T00:09:47.810289Z",{"id":266,"title":267,"url":268,"summary":269,"summary_zh":8,"content":8,"source_name":270,"source_url":8,"published_at":271,"category":12,"cover_url":8,"hotness":13,"is_selected":14,"score":272,"score_detail":273,"sources":276,"tags":278,"primary_topic":8,"secondary_topics":281,"classification":282,"search_phrases":293,"slug":296,"view_count":21,"doi":8,"paper":297,"created_at":304},4646,"面向共存植物叶部病害多标签诊断的类别感知视觉-语言框架CA-CLIP(CA-CLIP for multi-label diagnosis of co-occurring plant leaf diseases: A class-aware vision-language framework)","https:\u002F\u002Fwww.sciencedirect.com\u002Fscience\u002Farticle\u002Fpii\u002FS2772375526002688","研究提出CA-CLIP类别感知视觉-语言框架用于症状混合叶片多标签诊断,构建CCPLD基准(5种作物、3400张合成图像、17种共存病害类型)。在CCPLD验证集子集准确率达0.8543、Macro-F1达0.9601;蒸馏至MobileNetV3(2.99M参数)子集准确率0.8066,自然背景外部评估mAP 0.5626。局限:数据集为受控合成,自然场景跨域迁移能力仍有限。","Smart Agricultural Technology 7.1 2026-10-08","2026-10-08T00:00:00Z",77,{"impact":134,"substance":18,"depth":17,"authority":65,"freshness":274,"relevant":21,"comment":275},8,"提出类别感知视觉-语言框架并构建多标签病害基准，方法新颖、指标扎实，但数据集为受控合成、跨域迁移有限，属细分领域技术进展。",[277],{"name":270,"url":268},[70,71,28,279,280],"植物病害诊断","多标签识别",[],{"status":32,"locked":14,"version":33,"method":34,"evidence":283},{"candidates":284,"agricultural_evidence":43,"reason":44},[285,289],{"code":38,"score":39,"supported":14,"title":286,"abstract":287,"tags":288},[],[],[71,28],{"code":97,"score":21,"supported":14,"title":290,"abstract":291,"tags":292},[],[],[70],[294,295],"CA-CLIP 植物叶部病害 多标签诊断","CCPLD 基准 共存病害","CA-CLIP植物叶部病害多标签诊断-4646",{"doi":8,"openalex_id":8,"authors":298,"venue":8,"cited_by_count":105,"oa_url":8,"card":299,"direction":120,"ingested_from":263},[],{"tldr":300,"method":301,"finding":302,"direction":120,"opportunity":303},"提出CA-CLIP视觉-语言框架，实现共存植物叶部病害的多标签诊断。","构建CCPLD合成基准（5作物、3400图、17类），用CLIP类感知框架并蒸馏","子集准确率0.8543、Macro-F1 0.9601，蒸馏后0.8066，但自然场景跨域mAP仅0","合成数据到自然场景的跨域迁移是空白，可研究域适应与真实田间多标签病害数据构建。","2026-10-11T00:09:47.623718Z",{"id":306,"title":307,"url":308,"summary":309,"summary_zh":8,"content":8,"source_name":310,"source_url":308,"published_at":311,"category":12,"cover_url":8,"hotness":13,"is_selected":14,"score":312,"score_detail":313,"sources":315,"tags":317,"primary_topic":320,"secondary_topics":325,"classification":326,"search_phrases":343,"slug":346,"view_count":347,"doi":348,"paper":349,"created_at":363},3876,"Agrivlm: integrating vision-language models into real-time remote sensing for smart agriculture","https:\u002F\u002Fdoi.org\u002F10.1007\u002Fs11276-026-04228-x","Agrivlm: integrating vision-language models into real-time remote sensing for smart agriculture。Wireless Networks","Wireless Networks","2026-09-30T00:00:00Z",78,{"impact":17,"substance":19,"depth":17,"authority":65,"freshness":137,"relevant":21,"comment":314},"将视觉语言模型引入农业实时遥感，方法新颖且具应用潜力，值得关注。",[316],{"name":310,"url":308},[70,71,318,28,319],"实时监测","遥感",{"code":321,"slug":322,"name":323,"path":324,"parent_code":78},"I3","remote-sensing","遥感与农情监测","\u002Ftopic\u002Fdigital-agriculture\u002Fremote-sensing",[],{"status":81,"locked":14,"version":33,"method":34,"evidence":327},{"candidates":328,"agricultural_evidence":43},[329,334,339],{"code":97,"score":64,"supported":43,"title":330,"abstract":332,"tags":333},[331],"smart agriculture",[331],[70],{"code":321,"score":64,"supported":43,"title":335,"abstract":337,"tags":338},[336],"remote sensing",[336],[319],{"code":38,"score":39,"supported":14,"title":340,"abstract":341,"tags":342},[],[],[71,28],[344,345],"Agrivlm 视觉语言模型 遥感","智慧农业 实时遥感","Agrivlm视觉语言模型遥感-3876",3,"10.1007\u002Fs11276-026-04228-x",{"doi":348,"openalex_id":350,"authors":351,"venue":310,"cited_by_count":105,"oa_url":8,"card":8,"direction":201,"ingested_from":122},"W7214888751",[352,354,357,359,361],{"name":353,"orcid":8},"Sasmita Padhy",{"name":355,"orcid":356},"Preetam Suman","https:\u002F\u002Forcid.org\u002F0000-0001-7607-1372",{"name":358,"orcid":8},"Naween Kumar",{"name":360,"orcid":8},"Akansha Singh",{"name":362,"orcid":8},"Krishna Kant Singh","2026-10-01T23:30:08.990045Z",{"id":365,"title":366,"url":367,"summary":368,"summary_zh":8,"content":369,"source_name":370,"source_url":8,"published_at":371,"category":372,"cover_url":8,"hotness":13,"is_selected":14,"score":272,"score_detail":373,"sources":375,"tags":377,"primary_topic":380,"secondary_topics":381,"classification":382,"search_phrases":393,"slug":396,"view_count":150,"doi":8,"paper":8,"created_at":397},3597,"中国农科院农业信息研究所提出层次知识引导的枸杞害虫识别框架——Engineering Applications of Artificial Intelligence在线发表","https:\u002F\u002Fpaper.sciencenet.cn\u002Fhtmlnews\u002F2026\u002F9\u002F571988.shtm","中国农业科学院农业信息研究所科学数据研究室科研团队近日提出一种层次知识引导的枸杞害虫识别框架，有效解决真实农业场景下样本稀缺与类间混淆的双重挑战，相关研究成果发表在《人工智能的工程应用》（Engineering Applications of Artificial Intelligence）上。研究团队依托长文本-图像检索模型，把专家撰写的长文本描述解析为粗、中、细三个语义层次，并与多尺度视觉特征实现精准跨模态对齐。同时设计知识引导的去混淆模块，通过构建文本和视觉双模态相似度图，明显降低容易混淆类别之间的预测误差。该研究得到国家重点研发计划、中国农业科学院科技创新工程等项目支持。","[![Image 2: 科学网新闻频道](https:\u002F\u002Fpaper.sciencenet.cn\u002Fimages\u002Fnews.jpg)](https:\u002F\u002Fnews.sciencenet.cn\u002F)\n\n[生命科学](https:\u002F\u002Fwww.sciencenet.cn\u002Flife\u002F) | [医学科学](https:\u002F\u002Fwww.sciencenet.cn\u002Fmedicine\u002F) | [化学科学](https:\u002F\u002Fwww.sciencenet.cn\u002Fchemistry\u002F) | [工程材料](https:\u002F\u002Fwww.sciencenet.cn\u002Fmaterial\u002F) | [信息科学](https:\u002F\u002Fwww.sciencenet.cn\u002Finformation\u002F) | [地球科学](https:\u002F\u002Fwww.sciencenet.cn\u002Fearth\u002F) | [数理科学](https:\u002F\u002Fwww.sciencenet.cn\u002Fmathematics\u002F) | [管理综合](https:\u002F\u002Fwww.sciencenet.cn\u002Fpolicy\u002F)[站内规定](https:\u002F\u002Fblog.sciencenet.cn\u002Fblog-45-1064777.html) | [手机版](https:\u002F\u002Fwap.sciencenet.cn\u002F)\n\n[首页](https:\u002F\u002Fwww.sciencenet.cn\u002F) | [新闻](https:\u002F\u002Fnews.sciencenet.cn\u002F) | [博客](https:\u002F\u002Fblog.sciencenet.cn\u002Fblog.php) | [院士](https:\u002F\u002Fnews.sciencenet.cn\u002Fys\u002F) | [人才](https:\u002F\u002Ftalent.sciencenet.cn\u002F) | [会议](https:\u002F\u002Fmeeting.sciencenet.cn\u002F) | [基金·项目](https:\u002F\u002Ffund.sciencenet.cn\u002F) | [论文](https:\u002F\u002Fpaper.sciencenet.cn\u002F) | [绘图](https:\u002F\u002Fvisual-web.sciencenet.cn\u002F) | [视频·直播](https:\u002F\u002Fblog.sciencenet.cn\u002Fblog.php?mod=video) | [小柯机器人](https:\u002F\u002Fpaper.sciencenet.cn\u002FAInews) | [医学科普](https:\u002F\u002Fyxkp.sciencenet.cn\u002F)\n\n作者：李晨 来源：中国科学报 发布时间：2026\u002F9\u002F23 11:21:15 选择字号：小 中 大 新方法可智能识别枸杞害虫 近日，中国农业科学院农业信息研究所科学数据研究室科研团队，提出一种层次知识引导的枸杞害虫识别框架，有效解决真实农业场景下样本稀缺与类间混淆的双重挑战。相关研究成果发表在《人工智能的工程应用》（_**Engineering Applications of Artificial Intelligence**_）上。 ![Image 3](https:\u002F\u002Frmtzx.sciencenet.cn\u002Fkxwsprint\u002F6ab2a040e4b078fce44b0f9a.jpg)层次知识引导的枸杞害虫识别流程示意图。中国农科院供图 枸杞是高价值药用植物，虫害会对它的果实品质和药效造成很大影响。传统识别方法效率低且难以实现规模化应用。现有深度学习模型，面对外观高度相似的害虫种类时，常因标注样本不够，而识别效果不理想。 针对上述难题，科研团队依托长文本-图像检索模型，把专家撰写的长文本描述解析为粗、中、细三个语义层次，并与多尺度视觉特征实现精准跨模态对齐。同时，设计知识引导的去混淆模块，通过构建文本和视觉双模态相似度图，明显降低容易混淆类别之间的预测误差。该研究为枸杞等特色作物的虫害智能监测与精准防控提供新的技术路径，也为将农业专家知识融入视觉—语言模型提供参考。 该研究得到国家重点研发计划、中国农业科学院科技创新工程等项目支持。 相关论文信息：https:\u002F\u002Fdoi.org\u002F10.1016\u002Fj.engappai.2026.115342 版权声明：凡本网注明“来源：中国科学报、科学网、科学新闻杂志”的所有作品，网站转载，请在正文上方注明来源和作者，且不得对内容作实质性改动；微信公众号、头条号等新媒体平台，转载请联系授权。邮箱：shouquan@stimes.cn。\n\n打印 发E-mail给：\n\n以下评论只代表网友个人观点，不代表科学网观点。\n\n[![Image 4](https:\u002F\u002Fpaper.sciencenet.cn\u002Fimages\u002Fnewcomm.gif)](https:\u002F\u002Fpaper.sciencenet.cn\u002Fhtml\u002Fcomment.aspx?id=571988)\n\n相关新闻 相关论文\n*   1\n*   [陈文新：用后半生走遍中国“寻根”](https:\u002F\u002Fpaper.sciencenet.cn\u002Fhtmlnews\u002F2026\u002F9\u002F571931.shtm)\n\n*   2\n*   [上海市自然科学基金（启明星计划）项目申报指南发布](https:\u002F\u002Fpaper.sciencenet.cn\u002Fhtmlnews\u002F2026\u002F9\u002F571928.shtm)\n\n*   3\n*   [从实验室到生产线 高校以应用型科研对接产业升级](https:\u002F\u002Fpaper.sciencenet.cn\u002Fhtmlnews\u002F2026\u002F9\u002F571927.shtm)\n\n*   4\n*   [重点新材料研发及应用国家科技重大专项项目启动申报](https:\u002F\u002Fpaper.sciencenet.cn\u002Fhtmlnews\u002F2026\u002F9\u002F571926.shtm)\n\n*   5\n*   [一线治疗失败！晚期胃癌二线治疗有了“新标杆”](https:\u002F\u002Fpaper.sciencenet.cn\u002Fhtmlnews\u002F2026\u002F9\u002F571925.shtm)\n\n*   6\n*   [新策略助力有机室温磷光寿命大幅提升](https:\u002F\u002Fpaper.sciencenet.cn\u002Fhtmlnews\u002F2026\u002F9\u002F571909.shtm)\n\n*   7\n*   [AI绘制固态储氢材料全景导航 | 专家访谈](https:\u002F\u002Fpaper.sciencenet.cn\u002Fhtmlnews\u002F2026\u002F9\u002F571894.shtm)\n\n*   8\n*   [基金委发布1个国家自然科学基金专项项目申请指南](https:\u002F\u002Fpaper.sciencenet.cn\u002Fhtmlnews\u002F2026\u002F9\u002F571893.shtm)\n\n*   1\n*   [智能柔性传感器可原位实时监测植物体内激素](https:\u002F\u002Fpaper.sciencenet.cn\u002Fhtmlpaper\u002F2026\u002F9\u002F2026923105644661155910.shtm)\n\n*   2\n*   [浙江大学刘明团队Angew.：利用“受挫堆积”调控多孔有机笼孔道结构](https:\u002F\u002Fpaper.sciencenet.cn\u002Fhtmlpaper\u002F2026\u002F9\u002F202692310372413155909.shtm)\n\n*   3\n*   [GCB | 华南植物园揭示长期高氮沉降驱动热带森林根系分泌物加速磷活化的机制](https:\u002F\u002Fpaper.sciencenet.cn\u002Fhtmlpaper\u002F2026\u002F9\u002F2026923103553943155908.shtm)\n\n*   4\n*   [FEC | 土壤氮磷比调控高寒草地物种共存](https:\u002F\u002Fpaper.sciencenet.cn\u002Fhtmlpaper\u002F2026\u002F9\u002F2026923103451833155907.shtm)\n\n*   5\n*   [前沿聚焦：为什么烤羊肉串香？合肥工大与渤海大学联合揭示“香气密码”](https:\u002F\u002Fpaper.sciencenet.cn\u002Fhtmlpaper\u002F2026\u002F9\u002F2026923103345287155906.shtm)\n\n*   6\n*   [前沿聚焦：福州大学程树英-王伟煌联合河北大学高青：沉淀物回收策略助力 CBD 法 Sb2S3太阳电池提效](https:\u002F\u002Fpaper.sciencenet.cn\u002Fhtmlpaper\u002F2026\u002F9\u002F2026923103218818155905.shtm)\n\n*   7\n*   [Angew. Chem.：本征近红外二区荧光纳米酶用于高分辨活体成像与肿瘤催化免疫治疗](https:\u002F\u002Fpaper.sciencenet.cn\u002Fhtmlpaper\u002F2026\u002F9\u002F2026923102125755155904.shtm)\n\n*   8\n*   [2型糖尿病不同临床亚型的分子异质性被揭示](https:\u002F\u002Fpaper.sciencenet.cn\u002Fhtmlpaper\u002F2026\u002F9\u002F20269239501868155903.shtm)\n\n[图片新闻](https:\u002F\u002Fpaper.sciencenet.cn\u002Fimgnews.aspx)\n\n[![Image 5: “鹏城首发星”成功入轨](https:\u002F\u002Fnews.sciencenet.cn\u002F\u002Fupload\u002Fnews\u002Fimages\u002F2026\u002F9\u002F20269231515219900.png)](https:\u002F\u002Fnews.sciencenet.cn\u002Fhttps:\u002F\u002Fnews.sciencenet.cn\u002Fhtmlnews\u002F2026\u002F9\u002F571953.shtm)[![Image 6: 研究揭示甲烷部分氧化制合成气驱动因素](https:\u002F\u002Fnews.sciencenet.cn\u002F\u002Fupload\u002Fnews\u002Fimages\u002F2026\u002F9\u002F2026917125225680.jpg)](https:\u002F\u002Fnews.sciencenet.cn\u002F\u002Fhtmlnews\u002F2026\u002F9\u002F571610.shtm)\n[![Image 7: 科学家揭示火星大气物质运输规律](https:\u002F\u002Fnews.sciencenet.cn\u002F\u002Fupload\u002Fnews\u002Fimages\u002F2026\u002F9\u002F2026915174236470.jpg)](https:\u002F\u002Fnews.sciencenet.cn\u002F\u002Fhtmlnews\u002F2026\u002F9\u002F571503.shtm)[![Image 8: 超导双层镍氧化物薄膜的三维电子结构](https:\u002F\u002Fnews.sciencenet.cn\u002F\u002Fupload\u002Fnews\u002Fimages\u002F2026\u002F9\u002F20269151415154110.png)](https:\u002F\u002Fnews.sciencenet.cn\u002Fhttps:\u002F\u002Fnews.sciencenet.cn\u002Fhtmlnews\u002F2026\u002F9\u002F571484.shtm)\n[>>更多](https:\u002F\u002Fpaper.sciencenet.cn\u002Fimgnews.aspx)\n\n[一周新闻排行](https:\u002F\u002Fpaper.sciencenet.cn\u002Fzphnews.aspx)\n*   1\n*   [诺贝尔基金会宣布增加2026年诺奖奖金](https:\u002F\u002Fnews.sciencenet.cn\u002F\u002Fhtmlnews\u002F2026\u002F9\u002F571724.shtm)\n\n*   2\n*   [直播回放｜新型铁基催化剂，为海水制氢提速](https:\u002F\u002Fnews.sciencenet.cn\u002Fhttps:\u002F\u002Fweibo.com\u002Fl\u002Fwblive\u002Fp\u002Fshow\u002F1022:2321325343837114990618)\n\n*   3\n*   [汪品先院士：年轻人应跳出人类中心论](https:\u002F\u002Fnews.sciencenet.cn\u002F\u002Fhtmlnews\u002F2026\u002F9\u002F571776.shtm)\n\n*   4\n*   [从清华毕业做科普10年，我如何对抗AI](https:\u002F\u002Fnews.sciencenet.cn\u002F\u002Fhtmlnews\u002F2026\u002F9\u002F571671.shtm)\n\n*   5\n*   [袁慧书：脊柱“一扫多查”，让AI成为影像医生得力助手](https:\u002F\u002Fnews.sciencenet.cn\u002F\u002Fhtmlnews\u002F2026\u002F9\u002F571771.shtm)\n\n*   6\n*   [直播回放|马丁院士：共同追寻碳原子的“旅行路线”](https:\u002F\u002Fnews.sciencenet.cn\u002Fhttps:\u002F\u002Fweibo.com\u002Fl\u002Fwblive\u002Fp\u002Fshow\u002F1022:2321325344508505620608)\n\n*   7\n*   [新模型引爆AI圈：不生成文本，专为机器做快速决策](https:\u002F\u002Fnews.sciencenet.cn\u002F\u002Fhtmlnews\u002F2026\u002F9\u002F571794.shtm)\n\n*   8\n*   [AI绘制固态储氢材料全景导航 | 专家访谈](https:\u002F\u002Fnews.sciencenet.cn\u002F\u002Fhtmlnews\u002F2026\u002F9\u002F571894.shtm)\n\n*   9\n*   [AI正自主制造下一代AI？](https:\u002F\u002Fnews.sciencenet.cn\u002F\u002Fhtmlnews\u002F2026\u002F9\u002F571716.shtm)\n\n*   10\n*   [一周热闻回顾（2026年9月20日）](https:\u002F\u002Fnews.sciencenet.cn\u002F\u002Fhtmlnews\u002F2026\u002F9\u002F571820.shtm)\n\n[编辑部推荐博文](http:\u002F\u002Fblog.sciencenet.cn\u002Fblog.php?mod=recommend)\n*   ![Image 9](https:\u002F\u002Fpaper.sciencenet.cn\u002Fimages\u002Ft11.gif)\n*   [封面文章 | 效应蛋白TseMt的毒性机制及递送通路](https:\u002F\u002Fblog.sciencenet.cn\u002Fblog-45-1553810.html)\n\n*   ![Image 10](https:\u002F\u002Fpaper.sciencenet.cn\u002Fimages\u002Ft11.gif)\n*   [科学网2026年8月十佳博文榜单公布！](https:\u002F\u002Fblog.sciencenet.cn\u002Fblog-45-1551789.html)\n\n*   ![Image 11](https:\u002F\u002Fpaper.sciencenet.cn\u002Fimages\u002Ft11.gif)\n*   [我爱路亚](https:\u002F\u002Fblog.sciencenet.cn\u002Fblog-1185605-1554044.html)\n\n*   ![Image 12](https:\u002F\u002Fpaper.sciencenet.cn\u002Fimages\u002Ft11.gif)\n*   [学者的中秋家国](https:\u002F\u002Fblog.sciencenet.cn\u002Fblog-575926-1554043.html)\n\n*   ![Image 13](https:\u002F\u002Fpaper.sciencenet.cn\u002Fimages\u002Ft11.gif)\n*   [CD1 如何\"看见\"脂质](https:\u002F\u002Fblog.sciencenet.cn\u002Fblog-41174-1554027.html)\n\n*   ![Image 14](https:\u002F\u002Fpaper.sciencenet.cn\u002Fimages\u002Ft11.gif)\n*   [师傅领进门，修行在个人](https:\u002F\u002Fblog.sciencenet.cn\u002Fblog-729911-1553761.html)\n\n*   [更多>>](https:\u002F\u002Fblog.sciencenet.cn\u002Fblog.php?mod=recommend)\n\n[关于我们](http:\u002F\u002Fwww.sciencenet.cn\u002Faboutus.aspx) | [网站声明](http:\u002F\u002Fwww.sciencenet.cn\u002Fshengming.aspx) | [服务条款](http:\u002F\u002Fwww.sciencenet.cn\u002Ftiaokuang.aspx) | [联系方式](http:\u002F\u002Fwww.sciencenet.cn\u002Fcontact.aspx) | [举报](https:\u002F\u002Fwww.sciencenet.cn\u002Fjubao.aspx) | [中国科学报社](https:\u002F\u002Fstimes.sciencenet.cn\u002F)  \n[京ICP备07017567号-12](http:\u002F\u002Fwww.beian.miit.gov.cn\u002F)互联网新闻信息服务许可证10120230008 京公网安备 11010802032783  \nCopyright @ 2007-2026 中国科学报社 All Rights Reserved  \n地址：北京市海淀区中关村南一条乙三号 电话：010-62580783","科学网（中国科学报）2026-09-23","2026-09-23T00:00:00Z","报道",{"impact":17,"substance":19,"depth":136,"authority":20,"freshness":274,"relevant":21,"comment":374},"国家级科研机构在农业AI顶刊发表的技术突破，方法新颖且直击样本稀缺与类间混淆痛点，对特色作物虫害智能监测有示范价值，值得进入每日精选。",[376],{"name":370,"url":367},[70,71,378,28,379],"病虫害识别","枸杞",{"code":38,"slug":75,"name":76,"path":77,"parent_code":78},[],{"status":81,"locked":14,"version":33,"method":34,"evidence":383},{"candidates":384,"agricultural_evidence":43},[385,389],{"code":38,"score":20,"supported":43,"title":386,"abstract":387,"tags":388},[87],[87],[71,378,28],{"code":97,"score":21,"supported":14,"title":390,"abstract":391,"tags":392},[],[],[70],[394,395],"中国农科院 农业信息研究所 枸杞害虫","层次知识引导 害虫识别 视觉语言模型","中国农科院农业信息研究所枸杞害虫-3597","2026-09-27T00:05:17.078840Z"]