[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"article-32598":3},{"id":4,"title":5,"content":6,"froms":7,"author":8,"addTime":9,"classId":10,"channelTitle":11,"channelSlug":12,"parentChannelId":13,"parentChannelTitle":14,"parentChannelSlug":15,"photo":8,"filePath":8,"imgUrl":16,"seo":17,"previous":20,"next":23,"hotList":26,"recommendList":58},32598,"机器智能分类植物并估算植被覆盖获突破","\u003Ccenter style=\"font-family: Simsun; font-size: 14px; white-space: normal; background-color: rgb(255, 255, 255);\">\u003Cimg src=\"\u002FUploads\u002Fueditor\u002Fimage\u002F20191126\u002F6371037693879267116297145.jpg\" title=\"6.jpg\" _src=\"\u002FUploads\u002Fueditor\u002Fimage\u002F20191126\u002F6371037693879267116297145.jpg\" alt=\"6.jpg\"\u002F>\u003C\u002Fcenter>\u003Ccenter style=\"font-family: Simsun; font-size: 14px; white-space: normal; background-color: rgb(255, 255, 255);\">&nbsp;位于榆树稀树草原的项目超级大样地。 王锋供图\u003C\u002Fcenter>\u003Cp style=\"font-family: Simsun; font-size: 14px; white-space: normal; background-color: rgb(255, 255, 255);\">\u003C\u002Fp>\u003Cp style=\"font-family: Simsun; font-size: 14px; white-space: normal; background-color: rgb(255, 255, 255);\">\u003C\u002Fp>\u003Cp style=\"font-family: Simsun; font-size: 14px; white-space: normal; background-color: rgb(255, 255, 255);\">\u003C\u002Fp>\u003Cp style=\"font-family: Simsun; font-size: 14px; white-space: normal; background-color: rgb(255, 255, 255);\">\u003C\u002Fp>\u003Cp style=\"font-family: Simsun; font-size: 14px; white-space: normal; background-color: rgb(255, 255, 255);\">\u003C\u002Fp>\u003Cp style=\"font-family: Simsun; font-size: 14px; white-space: normal; background-color: rgb(255, 255, 255);\">\u003C\u002Fp>\u003Cp style=\"font-family: Simsun; font-size: 14px; white-space: normal; background-color: rgb(255, 255, 255);\">\u003C\u002Fp>\u003Cp style=\"font-family: Simsun; font-size: 14px; white-space: normal; background-color: rgb(255, 255, 255);\">\u003C\u002Fp>\u003Cp style=\"font-family: Simsun; font-size: 14px; white-space: normal; background-color: rgb(255, 255, 255);\">\u003C\u002Fp>\u003Cp style=\"font-family: Simsun; font-size: 14px; white-space: normal; background-color: rgb(255, 255, 255);\">\u003C\u002Fp>\u003Cp style=\"font-family: Simsun; font-size: 14px; white-space: normal; background-color: rgb(255, 255, 255);\">\u003C\u002Fp>\u003Cp style=\"text-indent: 2em;\">本报记者 丁洪美\u003C\u002Fp>\u003Cp style=\"text-indent: 2em;\">国际著名期刊《农业森林气象》2019年7月刊发一篇题为《应用无人机和机器学习进行旱地生态系统植被分类与覆盖度计算》的研究论文，阐述了一项基于无人机影像自动分类荒漠植被并计算其覆盖度的新方法。新方法效率提高了1.1-2.9倍，人们可以在公众平台（www.uav-hirap.org）免费使用，并基于100公顷大样地开发了一种降低地面植被调查工作量的样方设计优化模型。这篇论文的通讯作者为中国林业科学研究院荒漠化研究所的王锋，主要合作单位包括南京林业大学、加拿大新不伦瑞克大学、南京农业大学、中国科学院植物所等。该研究由国家“十三五”重点研发专项、国家自然科学基金和中国林科院院所长基金提供主要资助。\u003C\u002Fp>\u003Cp style=\"text-indent: 2em;\">据王锋介绍，植被覆盖的变化是反映旱地生态系统健康状况的关键指标。由于干旱地区木本和草本植被在小尺度上常常呈集群、混合分布，因此即使米级分辨率的高分辨率卫星遥感影像也难以准确分辨出木本与草本植被。这个项目是在无人机遥感快速发展的基础上，研发的在景观尺度上精确监测温带疏林草原植被动态的新方法。项目依托位于内蒙古自治区正蓝旗浑善达克沙地榆树稀树草原长期定位观测大样地（1km×1km，42°57′53″N、115°57′30″E），利用无人机获取的空间分辨率为0.1米\u002F像元正射影像，基于机器学习算法“分类和回归树”自动分类温带疏林草原上的木本、草本植物和裸地，计算各类型植被覆盖度。\u003C\u002Fp>\u003Ccenter style=\"font-family: Simsun; font-size: 14px; white-space: normal; background-color: rgb(255, 255, 255);\">\u003Cimg src=\"\u002FUploads\u002Fueditor\u002Fimage\u002F20191126\u002F6371037694536493881673694.jpg\" title=\"7.jpg\" _src=\"\u002FUploads\u002Fueditor\u002Fimage\u002F20191126\u002F6371037694536493881673694.jpg\" alt=\"7.jpg\"\u002F>\u003C\u002Fcenter>\u003Ccenter style=\"font-family: Simsun; font-size: 14px; white-space: normal; background-color: rgb(255, 255, 255);\">项目超级大样地冬季植被状况。 王锋供图\u003C\u002Fcenter>\u003Cp>&nbsp;&nbsp;&nbsp;&nbsp;这项研究的结果主要体现在3个方面。\u003C\u002Fp>\u003Cp>&nbsp; &nbsp; 一是可以快速、准确地对温带疏林草原的木本和草本植被进行自动分类并计算其盖度。利用目视解译和大样地地面调查的植被数据(包含4832棵乔木，18798棵灌木)进行了验证，机器分类结果与目视解译结果对比，整体精度为77.6%，平均精度为76.3%，Kappa系数为0.64。机器分类计算结果与地面植被调查结果相比，在不同空间尺度上具有稳定的线性关系。\u003C\u002Fp>\u003Cp>&nbsp; &nbsp; 二是比常见的机器学习方法计算效率显著提高。新方法比常见的机器学习算法，如支持向量机(SVM)、随机森林(RF)、梯度递增决策树(GBDT)的计算效率分别高出166%、111%、290%。新算法已开发完成“高分辨率无人机影像在线分析平台”，供研究人员自由使用。\u003C\u002Fp>\u003Cp>&nbsp; &nbsp; 三是对生态学野外取样设计的启示：构建模型优化设计地面调查样方。在进行野外植被覆盖度调查时，如何优化设计野外植被调查样方的大小与数量，使采样调查结果与整体植被覆盖度的误差满足需求，一直是生态学野外调查工作的难点。研究人员利用蒙特卡洛随机模拟方法模拟了取样样方大小、数量与总体植被覆盖度的偏差，发现了偏差与样方边长、数量遵循的函数关系。\u003C\u002Fp>\u003Cp>&nbsp; &nbsp; 中国林科院荒漠化研究所首席专家卢琦介绍，这项成果是中国林科院荒漠化研究所和中科院植物所于2013年联合共建的100公顷大样地取得的阶段性成果之一。无人机和人工智能技术将成为林业、生态学研究和生态环境质量监测的有力工具，可大大减少野外工时，提高地面调整效率。\u003C\u002Fp>\u003Ccenter style=\"font-family: Simsun; font-size: 14px; white-space: normal; background-color: rgb(255, 255, 255);\">\u003Cstrong>\u003Cspan style=\"font-size: 16pt;\">不同调查方法分类结果比较\u003C\u002Fspan>\u003C\u002Fstrong>\u003C\u002Fcenter>\u003Ccenter style=\"font-family: Simsun; font-size: 14px; white-space: normal; background-color: rgb(255, 255, 255);\">\u003Cimg src=\"\u002FUploads\u002Fueditor\u002Fimage\u002F20191126\u002F6371037695400751328179435.jpg\" title=\"8.jpg\" _src=\"\u002FUploads\u002Fueditor\u002Fimage\u002F20191126\u002F6371037695400751328179435.jpg\" alt=\"8.jpg\"\u002F>\u003C\u002Fcenter>\u003Ccenter style=\"font-family: Simsun; font-size: 14px; white-space: normal; background-color: rgb(255, 255, 255);\">&nbsp;&nbsp; 无人机航拍图&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;目视解译结果&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 机器学习分类结果图\u003C\u002Fcenter>\u003Ccenter style=\"font-family: Simsun; font-size: 14px; white-space: normal; background-color: rgb(255, 255, 255);\">&nbsp;\u003Cspan style=\"font-size: 16pt;\">\u003Cstrong>不同机器学习方法分类结果的视觉对比\u003C\u002Fstrong>\u003C\u002Fspan>\u003C\u002Fcenter>\u003Ccenter style=\"font-family: Simsun; font-size: 14px; white-space: normal; background-color: rgb(255, 255, 255);\">\u003Cimg src=\"\u002FUploads\u002Fueditor\u002Fimage\u002F20191126\u002F6371037696028292994998849.jpg\" title=\"9.jpg\" _src=\"\u002FUploads\u002Fueditor\u002Fimage\u002F20191126\u002F6371037696028292994998849.jpg\" alt=\"9.jpg\"\u002F>　\u003C\u002Fcenter>\u003Cp>\u003C\u002Fp>\u003Ccenter style=\"font-family: Simsun; font-size: 14px; white-space: normal; background-color: rgb(255, 255, 255);\">类和回归树模型(CART)&nbsp; &nbsp;&nbsp; &nbsp;&nbsp;支持向量机(SVM)&nbsp;&nbsp; &nbsp;&nbsp; &nbsp;&nbsp;随机森林(RF)&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; &nbsp;&nbsp;梯度递增决策树(GBDT）\u003C\u002Fcenter>\u003Cp style=\"text-indent: 2em;\">作者： 丁洪美\u003C\u002Fp>\u003Cp>\u003C\u002Fp>\u003Cp>\u003Cbr\u002F>\u003C\u002Fp>","中国绿色时报","","2019-11-26 14:52:41",31,"科研","research",5,"产业","industry","\u002FUploads\u002F20191126040133_484493.jpg",{"title":18,"keywords":5,"description":19},"机器智能分类植物并估算植被覆盖获突破 - 生态中国网","国际著名期刊《农业森林气象》2019年7月刊发一篇题为《应用无人机和机器学习进行旱地生态系统植被分类与覆盖度计算》的研究论文，阐述了一项基于无人机影像自动分类荒漠植被并计算其覆盖度的新方法。新方法效率提高了1.1-2.9倍，人们可以在公众平台（www.uav-hirap.org）免费使用，并基于100公顷大样地开发了一种降低地面植被调查工作量的样方设计优化模型。",{"id":21,"title":22},32636,"“森防管家”落地湖南张家界",{"id":24,"title":25},32487,"四获科技大奖，让中国松香走向世界——访中国工程院院士宋湛谦",[27,31,34,37,40,43,46,49,52,55],{"id":28,"title":29,"intro":30,"froms":30,"author":30,"addTime":30,"imgUrl":30,"indexImgUrl":30,"bigImgUrl":30,"topImgUrl":30,"lImgUrl":30,"filePath":30,"classId":30,"photo":30,"click":30},66335,"新能源汽车出海转向规模化常态化",null,{"id":32,"title":33,"intro":30,"froms":30,"author":30,"addTime":30,"imgUrl":30,"indexImgUrl":30,"bigImgUrl":30,"topImgUrl":30,"lImgUrl":30,"filePath":30,"classId":30,"photo":30,"click":30},66200,"中国清凉好物趁“热”出海    ",{"id":35,"title":36,"intro":30,"froms":30,"author":30,"addTime":30,"imgUrl":30,"indexImgUrl":30,"bigImgUrl":30,"topImgUrl":30,"lImgUrl":30,"filePath":30,"classId":30,"photo":30,"click":30},66188,"持续增强新兴产业发展确定性   ",{"id":38,"title":39,"intro":30,"froms":30,"author":30,"addTime":30,"imgUrl":30,"indexImgUrl":30,"bigImgUrl":30,"topImgUrl":30,"lImgUrl":30,"filePath":30,"classId":30,"photo":30,"click":30},66092,"新技术让废弃秸秆变身生物基塑料",{"id":41,"title":42,"intro":30,"froms":30,"author":30,"addTime":30,"imgUrl":30,"indexImgUrl":30,"bigImgUrl":30,"topImgUrl":30,"lImgUrl":30,"filePath":30,"classId":30,"photo":30,"click":30},66022,"电动汽车电池新国标落地    　　",{"id":44,"title":45,"intro":30,"froms":30,"author":30,"addTime":30,"imgUrl":30,"indexImgUrl":30,"bigImgUrl":30,"topImgUrl":30,"lImgUrl":30,"filePath":30,"classId":30,"photo":30,"click":30},65990,"全球极端热应激现象正不断加剧    ",{"id":47,"title":48,"intro":30,"froms":30,"author":30,"addTime":30,"imgUrl":30,"indexImgUrl":30,"bigImgUrl":30,"topImgUrl":30,"lImgUrl":30,"filePath":30,"classId":30,"photo":30,"click":30},65938,"建筑业的绿色转型，智能建造能做什么？ ",{"id":50,"title":51,"intro":30,"froms":30,"author":30,"addTime":30,"imgUrl":30,"indexImgUrl":30,"bigImgUrl":30,"topImgUrl":30,"lImgUrl":30,"filePath":30,"classId":30,"photo":30,"click":30},65901,"新一代智能制造向纵深发展    ",{"id":53,"title":54,"intro":30,"froms":30,"author":30,"addTime":30,"imgUrl":30,"indexImgUrl":30,"bigImgUrl":30,"topImgUrl":30,"lImgUrl":30,"filePath":30,"classId":30,"photo":30,"click":30},65841,"国外在推进地方应对气候变化立法方面有哪些经验？对我国有哪些启示？",{"id":56,"title":57,"intro":30,"froms":30,"author":30,"addTime":30,"imgUrl":30,"indexImgUrl":30,"bigImgUrl":30,"topImgUrl":30,"lImgUrl":30,"filePath":30,"classId":30,"photo":30,"click":30},65798,"国家能源局：  多措并举提升新能源并网消纳水平  ",[59,63,67,71,75],{"id":28,"title":29,"intro":30,"froms":30,"author":30,"addTime":30,"imgUrl":60,"indexImgUrl":61,"bigImgUrl":62,"topImgUrl":30,"lImgUrl":30,"filePath":30,"classId":30,"photo":30,"click":30},"\u002FUploads\u002FimageMovie\u002F20260812\u002Fe5164313dd09e3ad296ccc62216be575-1.png","\u002FUploads\u002FimageMovie\u002F20260812\u002Fe5164313dd09e3ad296ccc62216be575-2.png","\u002FUploads\u002FimageMovie\u002F20260812\u002Fe5164313dd09e3ad296ccc62216be575-3.png",{"id":32,"title":33,"intro":30,"froms":30,"author":30,"addTime":30,"imgUrl":64,"indexImgUrl":65,"bigImgUrl":66,"topImgUrl":30,"lImgUrl":30,"filePath":30,"classId":30,"photo":30,"click":30},"\u002FUploads\u002FimageMovie\u002F20260724\u002F9e6e1995504b3e8b073f49a14515888d-1.png","\u002FUploads\u002FimageMovie\u002F20260724\u002F9e6e1995504b3e8b073f49a14515888d-2.png","\u002FUploads\u002FimageMovie\u002F20260724\u002F9e6e1995504b3e8b073f49a14515888d-3.png",{"id":35,"title":36,"intro":30,"froms":30,"author":30,"addTime":30,"imgUrl":68,"indexImgUrl":69,"bigImgUrl":70,"topImgUrl":30,"lImgUrl":30,"filePath":30,"classId":30,"photo":30,"click":30},"\u002FUploads\u002FimageMovie\u002F20260723\u002F9c0d342dae7253a26417a36d262ef983-1.png","\u002FUploads\u002FimageMovie\u002F20260723\u002F9c0d342dae7253a26417a36d262ef983-2.png","\u002FUploads\u002FimageMovie\u002F20260723\u002F9c0d342dae7253a26417a36d262ef983-3.png",{"id":38,"title":39,"intro":30,"froms":30,"author":30,"addTime":30,"imgUrl":72,"indexImgUrl":73,"bigImgUrl":74,"topImgUrl":30,"lImgUrl":30,"filePath":30,"classId":30,"photo":30,"click":30},"\u002FUploads\u002FimageMovie\u002F20260713\u002Fa77c306b2b41f3a2ea8081a60b9b8b24-1.png","\u002FUploads\u002FimageMovie\u002F20260713\u002Fa77c306b2b41f3a2ea8081a60b9b8b24-2.png","\u002FUploads\u002FimageMovie\u002F20260713\u002Fa77c306b2b41f3a2ea8081a60b9b8b24-3.png",{"id":41,"title":42,"intro":30,"froms":30,"author":30,"addTime":30,"imgUrl":76,"indexImgUrl":77,"bigImgUrl":78,"topImgUrl":30,"lImgUrl":30,"filePath":30,"classId":30,"photo":30,"click":30},"\u002FUploads\u002FimageMovie\u002F20260707\u002Fc083b21084e4c3a01e3f8db30fbdb6a0-1.png","\u002FUploads\u002FimageMovie\u002F20260707\u002Fc083b21084e4c3a01e3f8db30fbdb6a0-2.png","\u002FUploads\u002FimageMovie\u002F20260707\u002Fc083b21084e4c3a01e3f8db30fbdb6a0-3.png"]