[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-2851":3},{"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":22,"tags":24,"view_count":30,"doi":31,"paper":32,"created_at":41},2851,"F2DMAS:基于智能手机视频的复杂背景盆栽植物三维表型工作流——Frontiers in Plant Science","https:\u002F\u002Fwww.frontiersin.org\u002Fjournals\u002Fplant-science\u002Farticles\u002F10.3389\u002Ffpls.2026.1885579\u002Ffull","F2DMAS将消费级智能手机视频、频域清晰度筛选、基于视觉基础模型的FSAM3植物前景提取、SfM位姿估计、2D Gaussian Splatting(2DGS)、TSDF显式网格、尺度恢复与表型测量串成一条连续流程。在复杂背景下,FSAM3取得F1-score 98.3%、mIoU 97.9%,HD95由SEEM的281.9 px降至41.4 px。F2DMAS达到PSNR 31.09 dB、SSIM 0.9711、LPIPS 0.0365;训练时间下降60.94%,网格提取时间下降65.17%。植株高度、冠幅、叶长、叶宽的R²分别约0.99、0.99、0.97、0.90;RMSE为1.21、0.99、0.64、0.64 cm。来自中国农业科学院农业信息研究所等单位。",null,"Frontiers in Plant Science","2026-09-17T00:00:00Z","论文",10,false,87,{"impact":16,"substance":17,"depth":18,"authority":19,"freshness":12,"relevant":20,"comment":21},22,23,18,14,1,"中国农科院信息所提出基于手机视频的盆栽植物三维表型工作流，精度与效率显著提升，方法新颖、数据扎实，值得进入每日精选。",[23],{"name":9,"url":6},[25,26,27,28,29],"智慧农业","农业人工智能","植物表型","三维重建","智能手机",0,"10.3389\u002Ffpls.2026.1885579\u002Ffull",{"doi":31,"openalex_id":8,"authors":33,"venue":8,"cited_by_count":30,"oa_url":8,"card":34,"direction":38,"ingested_from":40},[],{"tldr":35,"method":36,"finding":37,"direction":38,"opportunity":39},"提出基于智能手机视频的盆栽植物三维表型工作流F2DMAS，实现复杂背景下高效高精度重建与测量。","智能手机视频、频域清晰度筛选、FSAM3前景提取、SfM、2DGS、TSDF网格","前景提取F1达98.3%，重建PSNR 31.09 dB，株高冠幅R²约0.99，训练与网格提取时间","农业遥感与作物表型","可探索将该工作流迁移至田间自然场景、多物种及动态生长监测，并融合时序表型分析。","agent","2026-09-18T00:03:30.495696Z"]