无人机地形建模误差空间分布影响因素研究  

Study on the Influencing Factors of Spatial Distribution of Errors in UAV-Based Terrain Modeling

作　　者：卢旺达 王春 代文[1,4] 钱伟 李敏 位宏 杨昕 Lu Wangda;Wang Chun;Dai Wen;Qian Wei;Li Min;Wei Hong;Yang Xin(School of Remote Sensing and Geomatics Engineering,Nanjing University of Information Science and Technology,Nanjing 211800,China;Anhui Province Key Laboratory of Physical Geographical Environment,Chuzhou 239000,China;School of Geographic Information and Tourism,Chuzhou University,Chuzhou 239000,China;School of Geographical Sciences,Nanjing University of Information Science and Technology,Nanjing 211800,China;School of Geography,Nanjing Normal University,Nanjing 210023,China)

机构地区：[1]南京信息工程大学遥感与测绘工程学院,江苏南京211800 [2]实景地理环境安徽省重点实验室,安徽滁州239000 [3]滁州学院地理信息与旅游学院,安徽滁州239000 [4]南京信息工程大学地理科学学院,江苏南京211800 [5]南京师范大学地理科学学院,江苏南京210023

出　　处：《南京师大学报（自然科学版）》2023年第2期44-53,共10页Journal of Nanjing Normal University(Natural Science Edition)

基　　金：国家自然科学基金项目(42171402、41930102);安徽高校省级自然科学研究重大项目(KJ2021ZD0130);江苏省高等学校自然科学研究项目(22KJB170016);实景地理环境智能科技滁州市“113”产业创新团队、实景地理环境安徽省重点实验室开放课题(2022PGE013);南京信息工程大学科研启动经费(2022r019).

摘　　要：无人机发展迅速,已经在地球科学领域得到了广泛应用.前人以中误差(root mean square error,RMSE)为精度评价指标对影响无人机摄影测量精度的各类因素进行了大量研究.但是,基于无人机摄影测量的地形建模误差往往空间上变化分布,中误差无法反映误差的空间分布特征.因此,本文从误差空间分布的视角出发,通过计算误差空间分布图、误差的莫兰指数、样区整体的平均误差和标准误差,分析了相机倾角、航高和控制点数量对地形建模高程误差的大小及空间分布的影响.在黄土高原两个小流域的实验结果表明:(1)在无控制测量的情况下,误差受相机倾角的影响较大,采用较大角度的倾斜摄影不仅可以降低整体误差,还能改善误差的空间分布,减少误差的空间自相关性.(2)航高方面,尽管航高(60~160 m)变高会增大误差,但是航高对误差的空间分布影响不大.(3)在有控制测量的情况下,控制点的使用不仅降低了整体误差也优化了误差空间分布.在整体误差方面,使用少量的控制点即能达到一个稳定的精度水平.但此时,误差的空间分布还可以继续优化,要使样区的误差空间分布达到稳定的水平,需要相对较多的控制点.本研究为使用消费级无人机进行地形建模提供了有益的参考,在实际应用中可根据本文的结论优化航线设计方案和控制点布设.With the rapid development of UAVs,it has been widely used in the field of geoscience.A large number of previous studies have analyzed the various factors that affect the accuracy of UAV photogrammetry with RMSE as the accuracy evaluation index.However,terrain modeling errors based on UAV photogrammetry are often distributed in space,and RMSE cannot reflect the spatial distribution characteristics of errors.Thus,from the perspective of the spatial distribution of errors,this paper analyzes the influence of camera angle,flight height and the number of ground control points on the elevation error and spatial distribution of terrain modeling by calculating spatial distribution of errors,Moran’s Index of error,mean error and standard deviation error.Two small watershed areas on the Loess Plateau are selected as the study area.The experimental results show that:(1)Without ground control points,the error is greatly affect by the camera angle.Using higher oblique camera angles(higher than 10°)can not only reduce the overall error,but also improve the spatial distribution of errors and reduce the spatial autocorrelation of the error.(2)In the terms of flight height,although the increase of flight height(60-160 m)will increase the error,flight height has little effect on the spatial distribution of errors.(3)When there are GCPs,the use of GCPs reduces the overall error and optimizes the spatial distribution of errors.In the terms of overall error,a stable accuracy level can be achieved by using only a small number of GCPs.However,at this time,the spatial distribution of errors can continue to be optimized.More GCPs are required to make the spatial distribution of errors reach a stable level.This study provides a useful reference for terrain modeling using consumer UAV.In practical application,the measurement scheme and control points layout can be optimized according to the conclusions of this paper.

关 键 词：消费级无人机 地形建模 图像采集策略 误差空间分布 控制点数量 

分 类 号：P231[天文地球—摄影测量与遥感]