地面摄影RTK辅助SFM算法的影像定向  

Image Orientation Using SFM Algorithm Assisted by Ground Photography RTK

作　　者：邓非[1] 杨鹭漫 DENG Fei;YANG Luman(School of Geodesy and Geomatics,Wuhan University,Wuhan 430079,China)

机构地区：[1]武汉大学测绘学院,湖北武汉430079

出　　处：《测绘地理信息》2021年第5期8-12,共5页Journal of Geomatics

基　　金：国土资源部城市土地资源监测与仿真重点实验室开放研究基金(KF-2018-03-025)。

摘　　要：基于运动恢复结构(structure from motion,SFM)算法的地面影像定向通常没有考虑影像的位置姿态信息,影像定向结果在自由网坐标系,因此无法获取绝对坐标和真实尺度;利用地面摄影实时动态(real-time kinematic,RTK)技术可以获取地面影像的位置姿态信息,但是RTK和摄影测量的角度系统定义不同,RTK相位中心与影像投影中心存在偏差。针对这些问题,推导了利用RTK获取的GPS和角元素heading pitch roll(HPR)计算影像外方位元素的公式,提出地面摄影RTK辅助SFM算法的地面影像定向方法,在少量或者无需像控点条件下将SFM算法自由网的定向结果转换到绝对坐标系下,获得场景的真实坐标和尺度。实验结果表明,该转换公式正确,地面摄影RTK辅助SFM算法的影像定向方法可行。Ground image orientation based on structure from motion(SFM)algorithm usually does not consider the image information of position and attitude. The image orientation result is in a free coordinate system,so we can’t obtain absolute coordinate and true scale. Real-time kinematic(RTK)technology can be used to acquire position and attitude of ground image,however the angle system definitions of RTK and photogrammetry are different and there is a deviation between the RTK phase center and the image projection center.In response to these problems,we deduce the formula for calculating the external orientation elements of the image using the GPS and heading pitch roll(HPR)obtained by RTK technology. And we propose a method for image orientation using SFM algorithm assisted by ground photography RTK. Under the condition of few or no image control points,the free network orientation result of SFM algorithm is converted to the absolute coordinate system to obtain the real coordinates and scale of the scene. The experimental results show that the conversion formula is correct,and the method for image orientation using SFM algorithm assisted by ground photography RTK is feasible.

关 键 词：地面摄影实时动态(real-time kinematic RTK) 运动恢复结构(structure from motion SFM)算法 GPS/HPR 外方位元素 影像定向 

分 类 号：P228[天文地球—大地测量学与测量工程] TP751[天文地球—测绘科学与技术]