UAV SfM技术在活动构造研究中的应用——以青藏高原西北部龙木错断裂为例  

Application of UAV SfM technology in active tectonic research:A case study of the Longmu Co Fault,Northwestern Qinghai-Tibet Plateau

作　　者：江晨轶 潘家伟[1,2,3,4] 张丽军 李海兵 孙知明[6] Marie-Luce Chevalier 刘富财[1,2] 苏强 JIANG Chenyi;PAN Jiawei;ZHANG Lijun;LI Haibing;SUN Zhiming;CHEVALIER Marie-Luce;LIU Fucai;SU Qiang(Institute of Geology,Chinese Academy of Geological Sciences,Beijing 100037,China;Key Laboratory of Continental Dynamics of Ministry of Natural Resources,Beijing 100037,China;China Jiangsu Donghai Continental Deep Borehole Crustal Activity National Observation and Research Station,Lianyungang 222300,Jiangsu,China;SinoProbe Laboratory,Chinese Academy of Geological Sciences,Beijing 100094,China;Department of Hydraulic Engineering,Hebei University of Water Resources and Electric Engineering,Cangzhou 061001,Hebei,China;Institute of Geomechanics,Chinese Academy of Geological Sciences,Beijing 100081,China)

机构地区：[1]中国地质科学院地质研究所,北京100037 [2]自然资源部大陆动力学重点实验室,北京100037 [3]江苏东海大陆深孔地壳活动国家野外科学观测研究站,江苏连云港222300 [4]自然资源部深地科学与探测技术实验室,北京100094 [5]河北水利电力学院水利工程系,河北沧州061001 [6]中国地质科学院地质力学研究所,北京100081

出　　处：《地质力学学报》2024年第2期332-347,共16页Journal of Geomechanics

基　　金：中国地质调查局地质调查项目(DD20221630);国家自然科学基金项目(42372274);科技部科技基础资源调查专项(2021FY100101)。

摘　　要：为探讨搭载与未搭载实时动态差分技术/动态后处理技术(RTK/PPK)模块的无人机平台通过运动恢复结构(SfM)方法处理获得的数字高程模型(DEM)数据质量差异,以及建立不同无人机平台野外数据采集和室内数据处理过程的快速流程,利用大疆经纬M 300 RTK无人机(搭载禅思L1激光雷达(LiDAR)和测绘相机)与大疆精灵4 Pro无人机(搭载可见光相机)分别对青藏高原西北部龙木错断裂上1处位错阶地进行了数据采集,获得了该处高分辨率、高精度DEM数据和数字正射影像图(DOM)数据。对比结果显示,M 300 RTK无人机平台L1负载系统获得的LiDAR和SfM地形数据精度接近,两者在约100 m的飞行高度获得的DEM数据在水平和垂直方向的均方差分别为0.135 m、0.111 m和0.201 m、0.180 m;无RTK模块的精灵4 Pro无人机获取的DEM数据虽然绝对精度较差(水平和垂直方向均方差分别为1.707 m和249.280 m),但其反映的相对地形与实际地形接近,经过地面控制点校正后精度可以达到分米级。研究表明,RTK SfM技术克服了使用地面控制点的局限性,为活动构造研究领域微地貌测量提供了更高精度、更高效率的解决方案。当对测区的绝对三维坐标要求不高,仅需相对的地形起伏时,未搭载RTK模块的无人机也能够在无地面控制点约束的情况下满足地貌位错测量基本需求。[Objective]In recent years,UAV(Unmanned Aerial Vehicle)SfM(Structure from Motion)technology has been widely applied as an emerging high-precision 3D topographic data acquisition technique in active tectonics research.However,existing domestic research primarily focuses on UAV platforms without RTK/PPK modules.Over the past two years,RTK and PPK technologies have gradually been introduced to UAV platforms,resulting in changes to field data collection and in-office data processing workflows compared to traditional UAV platforms without RTK/PPK modules.The differences in topographic data quality obtained through SfM processing of aerial photographs from UAV platforms equipped and not equipped with RTK/PPK modules under the same scenes and collection conditions still require investigation.[Methods]To address these questions and establish a streamlined data collection and processing workflow for different UAV platforms,a site where alluvial terraces are offset by the sinistral Longmu Co fault,northwestern Qinghai-Tibet Plateau,was surveyed using the DJI M300 RTK UAV(equipped with the Zenmuse L1 LiDAR and a 20-megapixel visible light camera)and the DJI Phantom 4 Pro UAV(equipped with a 20-megapixel visible light camera).Through processing the raw data with SfM technology,high-resolution and high-precision DEM and DOM data were obtained for the study area.Additionally,16 ground control points(checkpoints)were uniformly distributed in the research area,and their coordinates were measured using the Trimble R8 GNSS receiver in RTK mode to compare and validate the differences in data quality obtained by the two platforms.[Results]The data comparison showed that the visible light camera and LiDAR module carried by the M300 RTK demonstrated high accuracy,with root mean square errors(RMSEs)in the centimeter to decimeter range compared to the RTK-measured ground checkpoints.Compared to SfM data collected at the same time(RMSEX=0.176,RMSEY=0.099,RMSEZ=0.180,RMSEH=0.201,RMSE3D=0.270,unit:m),the Zenmuse L1 LiDAR data exhibited sli

关 键 词：无人机 运动恢复结构 摄影测量 活动构造 精度评估 龙木错断裂 

分 类 号：P231[天文地球—摄影测量与遥感] P65[天文地球—测绘科学与技术]