基于无人机SfM数据的挺水植物生物量反演  

Estimating biomass of emergent aquatic plants based on UAV SfM data

作　　者：井然[1,2,3] 宫兆宁[1,2,3] 赵文吉[1,2,3] 邓磊[1,2,3] 阿多[1,2,3] 孙伟东 

机构地区：[1]首都师范大学资源环境与旅游学院,北京100048 [2]三维信息获取与应用教育部重点实验室,北京100048 [3]资源环境与地理信息系统北京市重点实验室,北京100048

出　　处：《生态学报》2017年第22期7698-7709,共12页Acta Ecologica Sinica

基　　金：国家国际科技合作专项资助(2014DFA21620)

摘　　要：生物量是衡量挺水植物生长状况的重要参数,对湿地生态系统健康评价具有重要意义。利用无人机影像生成运动重建结构Sf M(Structure from Motion,Sf M)数据,结合野外实测生物量构建定量反演模型,并根据反演模型对生物量进行空间制图,最后分析了挺水植物类型对生物量空间分布的影响。结果表明,文中基于Sf M数据建立的逐步线性回归模型(Stepwise Linear(SWL)regression model)具有较好的反演精度及估测能力。其模型显著性为显著(P<0.01),决定系数为0.86,相对均方根误差为6.1%。挺水植物类型对生物量空间分布影响显著(P<0.05)。通过对研究区挺水植物的生物量进行估算,为利用无人机遥感监测挺水植物生物量提供了新思路。Biomass is an important ecological parameter that is used to evaluate the growth condition of emergent plants in wetlands during ecosystem health assessments. This study used SfM （Structure from Motion） data generated from UAV images and field measurements of emergent plant biomass to establish a quantitative relationship between the SfM data and biomass, which was then used to map biomass in the study area. The influence of emergent plant types on the spatial distribution of biomass was analyzed. Our results show that a Stepwise Linear regression （SWL） model based on the SfM data had the best forecasting accuracy and ability （P〈 0.01 ）, with a coefficient of determination （R2） of 0.86 and an rRMSE of 6.1%. Emergent plant types had a significant influence （P〈0.05） on the spatial distribution of biomass in the study area. The results of this study provide a new quantitative method for retrieving growth parameters for emergent aquatic plants.

关 键 词：挺水植物 生物量 无人机影像 运动重建结构(SfM)数据 回归分析 

分 类 号：Q948[生物学—植物学]