无人机倾斜航空摄影监测崩岗侵蚀量变化的方法

如何高效精确地监测崩岗的动态发育过程并且量化侵蚀量是崩岗侵蚀机理研究中的难点。该文以准专业级无人机对目标崩岗进行倾斜摄影获得的全方位多角度航空影像为基础，通过空三加密处理生成目标崩岗的三维点云模型；利用点云数据构建DTM，提取目标崩岗地形数据；运用多时相连续DEM相减的方法获取监测周期内崩岗的高程变化，计算侵蚀量并找到侵蚀严重的部位，再使用2.5D体积测算方法细化侵蚀严重的崩壁和沟头部位的侵蚀量，以此作为补充，最终获得监测期内的总侵蚀/沉积量体积并换算为泥沙量。最终结果验证的平均相对误差为9.69%，一个月监测周期内最大的绝对误差仅为0.303 3 m3，满足监测要求。因此利用无人机倾斜航空摄影测量的方法监测崩岗侵蚀量是可行有效的，该方法可提取崩岗的所有地形信息，研究侵蚀泥沙的来源和侵蚀过程，是较为快速和精确的崩岗监测手段。

Method for monitoring change in Benggang erosion based on oblique aerial images of UAV

Abstract: Benggang erosion is one of the most serious ecological and environemental problems in southern China. Accurate monitoring of the dynamic development and quantification of the erosion amount are big challenges in the study of Benggang erosion mechanism, due to the complexity of landform and topography of Benggang erosional areas. The process of Benggang erosion is dominated by collapse, making the difficuities in eorison estimation by raditional monitoring method or by the high-altitude aerial photography, especially for the calculation of the collapse of the steep wall.Three-dimensional laser scanning, as an optical measurement in erosion estimation for its high accuracy, is time-consuming and labor-consuming, and is less applicable for monitoring in the field conditions due to the heavy weight of the instrument. Obviously, a more concise and accurate method is needed in the study of Benggang erosion mechanism. Unmanned aerial vehicle (UAV) oblique aerial photography technology with high-resolution images is an alternative approach for the generation of point cloud data and extraction of different Benggang terrain parameters by point cloud information, imporving the limitions of traditional measurement. In this paper, a representative Benggang was selected according to its landform specifications and flight operating conditions which is located in Tongcheng County, Hubei Province. A multi-rotor UAV (DJ Phantom 4 Pro.) equipped with 20 million pixel CMOS image sensors was used as data acquisition platform in the estimation of Benggang erosion. The multi-angle oblique aerial images of the Benggang were obtained by using the UAV in the manner of spiral flying. In this operation, the flight status of the UAV and the status of camera haeundae were controlled by the main remote controllor and the auxiliary remote controller, respectively. Point data of different seasonal periods of Benggang development were collectd by five flights of the UAV from June to September in 2018, and three-dimensional laser scanner was simultaneously operated for the valiadation of UAV. The collapse amount of Benggang erosion is calculated according to the following three steps:1) constructing DTM of the target Benggang by generating the three-dimensional point cloud model through emptying three encryption processing, and extracting target collapse terrain data; 2) collecting the elevation change of collapse during the monitoring cycles by multi-time continuous DEM subtraction method, calculating the erosion amount and identifying the most severely eroded site by the combination of lowest level method (LODmin)anddifference of DEMs (DOD) uncertainty threshold; 3) Refining the erosion amount of the most severly eroded wall and gully head by 2.5D volume calculation method, and the total erosion / deposition volume which could be converted into mud and sediment volume during the monitoring period. The results showed that: In comparsion with three-dimensional laser scanner, the relative error of the proposed method was 9.69% in average, and the maximum absolute error was only 0.3033 m3 in one month''s monitoring cycle. The results satisfied the monitoring requirements. Therefore, UAV tilt aerial photogrammetry is effective and efficient in monitoring the collapse amount of Benggang. The source and process of erosion sediment can be identified with all the topographic information of Benggang extracted by the proposed method.