国产机载LiDAR点云与航空影像融合获取高精度DEM技术研究

发布时间：2018-05-21 03:36

本文选题：国产LiDAR + 系统检校　；参考：《昆明理工大学》2017年硕士论文

【摘要】：机载激光雷达作为一种快速、实时、高效的获取地理空间信息的新型技术,对传统测绘事业的发展起到了巨大的推动作用。本文围绕着国产机载LiDAR系统AP-3500的定位模型建立,集成误差检校,获取的点云与影像融合等方面进行研究,以提高点云定位精度,并通过实验对所获取点云的精度进行验证。本文完成的主要工作如下:(1)系统的集成研究及误差修正模型建立。对系统的硬件设备组成及工作原理做了简要介绍,并阐述了惯性导航系统的定位原理与LiDAR点云的测量原理,同时,为了得到三维激光脚点精确的三维坐标,在系统各部件之间建立相对独立的坐标系,通过研究建立起了消除误差的数学模型,找到了各坐标系间的位置与角度偏差。(2)集成误差检校。建立起了解求单相机的畸变系数与内方外元素的检校方式,并完成了双拼相机的平台检校,且通过对激光扫描仪的测距测角误差进行检校,建立起了扫描仪的检校方案,同时完成了 POS与LiDAR系统及相机系统的集成误差检校,构建了一整套误差集成检校方法。(3)数据处理研究。通过实验,得到了对影像辐射信息损失相对最少,变形相对最小的影像处理流程与方法,且借助影像信息的辅助对点云数据进行处理,并成功将两种数据进行融合,使融合后的点云即拥有强度,回波次数,反射强度,三维坐标信息等原有信息,还具备了影像光谱信息,随后借助这些特征信息对点云进行过滤与分类处理,利用滤波所得的地面点构建地面三角网,生成高精度DEM。(4)DEM精度验证与分析。在国家测绘地理信息局相关科研项目的支撑下,在河南省平顶山市卫东区某地进行飞行试验,利用同机飞行获取的影像与点云进行融合操作,并利用融合后的综合信息进行过滤与分类处理,适当的加以手工干预后,利用处理所得地面点构建DEM,并利用在实验区放置的正方形标靶的实测中心与拟合中心的坐标差值来验证点云的平面与高程精度,最后使用检查点检测法对DEM进行精度评定。
[Abstract]:Airborne lidar, as a new technology to obtain geospatial information quickly, real-time and efficiently, has played a great role in promoting the development of traditional surveying and mapping. In order to improve the accuracy of point cloud location, this paper focuses on the establishment of the localization model of the domestic airborne LiDAR system AP-3500, the integration error checking, the fusion of the obtained point cloud and the image, and verifies the accuracy of the obtained point cloud through experiments. The main work of this paper is as follows: 1) Integration of the system and establishment of error correction model. The hardware components and working principle of the system are briefly introduced, and the positioning principle of inertial navigation system and the measuring principle of LiDAR point cloud are expounded. At the same time, in order to obtain the accurate 3D coordinate of 3D laser foot-point, A relatively independent coordinate system is established among the various parts of the system. A mathematical model for eliminating errors is established and the position and angle deviation between the coordinate systems is found. In this paper, the distortion coefficient of single camera and the checking mode of inner and outer elements are established, and the platform check of double camera is completed, and the calibration scheme of scanner is established by checking the ranging and angle error of laser scanner. At the same time, the integrated error checking of POS, LiDAR system and camera system is completed, and a set of error integrated calibration method. Through experiments, the image processing flow and method with the least loss of image radiation information and the smallest deformation are obtained, and the point cloud data is processed with the aid of image information, and the two kinds of data are fused successfully. The fused point cloud has the original information, such as intensity, echo times, reflection intensity, 3D coordinate information and so on. It also has the spectral information of image, and then filters and classifies the point cloud with the help of these characteristic information. The ground triangulation network is constructed by using the ground points obtained by filtering, and high precision DEM.(4)DEM accuracy verification and analysis are generated. Supported by relevant scientific research projects of the State Bureau of surveying and Mapping Geographic Information, a flight test was conducted in the Weidong District of Pingdingshan City, Henan Province. The image obtained from the same flight was fused with the point cloud. The integrated information was filtered and classified, and the appropriate manual intervention was carried out. DEM is constructed by using the treated ground points, and the coordinate difference between the measured center and the fitting center of the square target placed in the experimental area is used to verify the plane and elevation accuracy of the point cloud. Finally, the accuracy of DEM is evaluated by checkpoint detection method.
【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P23

【参考文献】