基于倾斜摄影的无人机高空摄影测量技术是近年来的重点研究方向，与传统测绘技术方法比较，它具有相应速度快、辨识度高、操控方便、成本低等诸多优势。不但可以获取物体的纹理和位置信息，而且解决了传统的影像无法获取的细节信息。随着计算机软硬件水平的不断提升，进一步推动了该项技术的迅速普及和不断发展。近年来，随着国家提出的数字地球、智慧城市建设、建筑物景观设计、智慧社区、智慧工厂等发展规划的提出，对高还原度的建筑物3D模型的需求更加迫切，这些需求都要通过无人机测绘的技术得以实现。因此基于无人机倾斜摄影方法的实景三维建模方案具有很好的研究背景和广泛的应用需求。

本文主要利用高空的多姿态飞行倾斜摄影方法，针对多种复杂地形详细探讨了各种地形的飞行方案与内业空三处理方案，并结合实际案例进行了综合阐述。以整个无人机倾斜摄影测量的三维建模技术流程为主线，研究分析了无人机倾斜摄影测量系统的软硬件构成、测绘流程以及建筑物三维模型重建过程的关键技术等，并采用无人机多姿态图片采集测量技术，针对实际应用案例实现了复杂地形的三维模型建立和成果精度分析。本文的主要研究内容为：

（1）研究分析了无人机在复杂地形下，三维建模的关键技术，重点对三维建模中的特征点提取，点云和多视影像匹配算法进行了深入研究，并对三维模型的主流校正算法进行了分析。

（2）以无人机在复杂地形三维建模过程中的主要影响因素和改进方法为研究重点，首先以实际飞行数据质量为依据，对比分析了像片控制点布设的不同网型和边长，不同像片重叠度，航线的规划与设计中不同航高和航线形状以及不同飞行起飞角度和建筑物排列方式等因素进行了深入探究。针对复杂地形地貌，列举了四种地形做了详细阐述，它们分别是测区中央有高凸起障碍地形、测区一端地势陡峭地形、两山脊夹一谷地形与测区多起伏地形。同时，给出了多种实际生产中，遇到的内业空三难题与其解决方案，包括空地融合技术与空三解算、多源数据融合构建精细化三维模型与大区域分区空三平差后联合平差。

（3）以实际的应用案例为实验数据，对实际测量条件下无人机飞行数据采集和作业成果生产技术的改进，通过最后实验成果三维模型的相关数据精度进行对比分析，表明最终建模精度在像控点布设方案、飞行数据采集、影像重叠度设置、三维建模算法等方面，不但可以满足测绘生产中数字测图规范的精度要求，而且项目整体精度有所提升。

UAV aerial photogrammetry technology based on tilt photography is a key research direction in recent years. Compared with the traditional surveying and mapping technology, it has many advantages, such as fast speed, high identification, easy operation and low cost. It can not only obtain the texture and position information of the object, but also solve the detail information which can not be obtained by the traditional image. With the continuous improvement of computer hardware and software level, it further promotes the rapid popularization and continuous development of this technology. This technology realizes the demand of building 3D model with high reduction degree, and with the digital earth, smart city construction, building landscape design, smart community, factory and other contents proposed by the state in recent years, all of them can be realized by the above UAV mapping technology. Therefore, the real scene 3D modeling scheme based on UAV tilt photography method is very good The research background and wide application needs of.

    This paper mainly uses the high altitude multi attitude flight tilt photography method, aiming at a variety of complex terrain, discusses in detail the flight scheme of various terrain and the three processing scheme of indoor work, and gives a comprehensive description combined with practical cases. Taking the whole UAV tilt photogrammetry 3D modeling technology process as the main line, this paper studies and analyzes the software and hardware composition of UAV tilt photogrammetry system, the mapping process and the key technology of building 3D model reconstruction process, and uses UAV multi pose image acquisition measurement technology to realize the 3D model establishment and accuracy of complex terrain for practical application cases analysis. The main contents of this paper are as follows:

(1) This paper studies and analyzes the key technologies of UAV 3D modeling in complex terrain, focuses on the feature point extraction, point cloud and multi view image matching algorithm in 3D modeling, and analyzes the mainstream correction algorithm of 3D model.

This paper focuses on the main influencing factors and improvement methods of UAV in the process of 3D modeling of complex terrain. Firstly, based on the quality of actual flight data, this paper compares and analyzes the different net type and side length of photo control points, different photo overlap degree, different flight height and route shape in route planning and design, as well as different flight take-off angle and building arrangement And so on. According to the complex topography, four kinds of topography are listed in detail, they are the high raised obstacle terrain in the center of the survey area, the steep terrain at one end of the survey area, the topography of two ridges with a valley and the multi undulating terrain in the survey area. At the same time, a variety of practical production problems and solutions are given, including air-to-ground fusion technology and air-to-air three-dimensional solution, multi-source data fusion to build a refined three-dimensional model and large-scale regional spatial three-dimensional adjustment after joint adjustm.

(3) Taking the actual application case as the experimental data, this paper improves the UAV flight data acquisition and production technology under the actual measurement conditions. Through the comparative analysis of the relevant data accuracy of the final three-dimensional model of the experimental results, it shows that the final modeling accuracy is in the aspects of image control point layout scheme, flight data acquisition, image overlap setting, 3D modeling algorithm, etc It can meet the accuracy requirements of digital mapping specification in surveying and mapping production, and the overall accuracy of the project has been improved.

[1]张祖勋.数字摄影测量与计算机视觉[J].武汉大学学报.信息科学版, 2004, 29(12): 1035-1105.

[2]王庆栋. 新型倾斜航空摄影技术在城市建模中的应用研究[D].兰州交通大学,2013.

[3]李天兰. 三维点云数据的处理与应用[D].昆明理工大学,2011.

[4]刘学，张弘.基于三维激光扫面数据的构筑物三角构网模型[J].测绘科学,2008,33(1):1-4.

[5]杜飞轮,杜秦川.未来30年我国资源环境趋势及促进绿色发展的建议[J]．中国经贸导刊:理论版，2018(3):15-18.

[6]曹琳. 基于无人机倾斜摄影测量技术的三维建模及其精度分析[D].西安科技大学,2016.

[7]范秀庆.无人机免像控技术在地形图测量中的应用[J].测绘通报,2017(1):159-160.

[8]黄莺.基于ORB和SIFT的特征提取算法[J].哈尔滨师范大学(自然科学学报).2017,(2):25-29.

[9]杨文环. 结合空三与SFM的近景摄影测量点云获取研究[D].中国矿业大学.2016.

[10]谢榛. 基于无人机航拍图像的室外场景三维重建技术研究[D].浙江工业大学,2014.

[11]李维江.农用无人机的研究现状及趋势[J]．农业与技术,2018(10):27-31.

[12]张云硕,史云天,董云哲等.农用无人机喷洒技术的研究[J].农业与技术,2015.35(21):46-47.

[13]温源,张向东,沈建文,等.中国无人机发展技术路线及行业趋势[J].农业技术与装备,2014(5):35-38.

[14]娄尚易,薛新宇,顾伟,等.农用无人机的研究现状及趋势[J].农机化研究,2017,39(12):1-6.

[15]曹国强,章敏.无人机需求分析与展望[J].农业与技术,2016,36(21):65-66.

[16]高井样,王坚,李增科.智能背景下测绘科技发展的几点思考[J].武汉大学学报.信息科学版,2019,44(1):55-61.

[17] R.Hartly, Stereo from uncalibrated cameras[A]. Proc. CVPR 92[C]. Urbana-Champaign, Illinois: IEEE Computer Society. 1992, 761-764.

[18] Polleleys M. Self-calibration and metric 3D reconstruction from uncalibrated image sequences[D], phD. Thesis, Katholieke Universiteit Leuven, Heverlee,1999.

[19] Noah snavely, Steven M. Seitz, Richard Szeliski. Modeling the World from Internet Photo Collections[J]. International Journal of Computer Vision November 2008, Volume 80, Issue 2, p 189-210.

[20] Yasutaka furukawa, Brian Curless, Steve M. Seitz, and Richard Szeliski. Towards Internet-scale Multi-view Stereo[J]. Computer Vision and Pattern Recognition, 2010.

[21] Yastutaka Furukawa and Jean Ponce Accurate, Dense, and Robust Multi-View Stereopsis[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol.32, Issue 8, Pages 1362-1376, August 2010.

[22] Qi Shan, Riley Adams, Brian Curless, Yasutaka, and Steven M. Seitz The Visual Turing Test for Scene Reconstruction [J]. International Conference on 3D Vision 3DV 2013.

[23]刘经南,高柯夫.智能时代测绘与位置服务领域的挑战与机遇[J].武汉大学学报.信息科学版,2017,42(11):1506-1517.

[24]邵振程. 应用于农业植保领域的无人机测绘技术研究[D].杭州电子科技大学,2019.

[25]罗怀波.新地理信息时代的信息化测绘分析[J]．南方农机,2018(4):72-75.

[26]王小平,王根铎.测绘无人机的智能化设计[J].测绘技术装备,2011(3):2011(3):60-61.

[27]王京卫.测绘无人机低空数字航摄影像去雾霾研究[J].测绘学报,2016,45(2):19-22.

[28]张莞玲.无人机低空摄影测晕系统的应用[J].工程技术研究,2017(4):25-30.

[29]李泉洲,石高峰,崔建甲.无人机测绘遥感[J].电信工程技术与标准化,2017,30(4):14-17.

[30]张涵.无人机在测绘工程中应用技术的分析[J].硅谷,2014(16):127-128.

[31]支卫斌.无人机航空摄影测量在地形测绘中的应用[J].江西建材,2015(8):224-225.

[32]赵吉先,聂运菊.测绘仪器发展的回顾与展望[J].测绘通报,2008(2):70-71.

[33]胡奎,王丽英.三维激光扫描技术在精细地形图绘制中的应用[J].辽宁工程技术大学学报(自然科学版),2013,32(7):953-956.

[34]毕凯,李英成,丁晓波,等.轻小型无人机航摄技术现状及发展趋势[J].测绘通报,2015(3):27-48.

[35]张庆英.迈向测绘地理信息事业新征程[J].中国测绘,2012(4):6-7.

[36]论地理国情监测的技术支撑[J].武汉大学学报(信息科学版),2012,37(5):505-512.

[37]张骥.测绘应急保障中航空影像获取的方法[J].测绘标准化,2015(1):39-40.

[38]孟样辉,李浩川,王冠珠,等.高分遥感监测与规划辅助决策系统设计与实现[J].电子技术与软件工程,2018(6):31-34.

[39]杨德芳,田晓程.浅谈青海省测绘应急保障体系建设[J].青海国土经济,2013(5).

[40]王晓莉.基于无人机低空遥感系统的矿区地形图快速测绘[J].地理空间信息,2013,11(6):41-43.

[41]杨冠雄. 基于全景视觉的运动目标检测方法研究[D].哈尔滨工程大学,2010.

[42]张毅. 地面三维激光扫描技术在龟山汉墓测量和重建中的应用[D].长安大学.2013.

[43]牛亚晓,张立元,韩文霆,等.基于无人机遥感与植被指数的冬小麦覆盖度提取方法[J].农业机械学报,2018(4):20-22.

[44]陈茹雯,陈自新,史金飞,等.基于数学模型的图像畸变校正方法研究[C].全国机械工程博士生学术论坛.2007.

[45]赫美琳,高明慧,李爽,等.一种单目相机标定算法研究[J].数字通信世界,2018(5):42-45.

[46]刘松林,韦华.基于基本矩阵的相机畸变差校正方法研究[J].遥感信息,2009(3):7-10.

[47]朱师欢.自动空三点特征高精度影像匹配研究[D].西安科技大学.2017.

[48]杨晓敏,吴炜,卿粼波,等.图像特征点提取及匹配技术[J].光学精密工程,2009,17(9):2276-2282.

[49]黄松梅,毕远伟,许晓.双目立体匹配算法的研究与实现[J].鲁东大学学报：自然科学版,2018(1):25-30.

[50]葛忠孝,耶帅,夏琴,等.基于树形结构的半全局立体匹配算法[J].计算机工程,2016,42(8):243-248.

[51]李二涛,张国煊,基于最小二乘的曲而拟合算法研究[J].杭州电子科技大学学报:自然科学版,2009,29(2):48-51.

[52]陈宇,白征东.基于非线性最小二乘法的空间坐标转换[J].大地测量与地球动力学,2010,30(2):129-132.

[53]金泽选,吴开华,王朔,等.全自动无人机含障不规则区域的航迹规划算法[J].江苏农业科学,2018,46(5):220-223.

[54]陈建龙,杨德明,华庆海.实时动态(RTK)定位技术在土地测绘中的应用[J].测绘与空间地理信息,2000,23(3):20-21.

[55]方诗明,姬海莉.航测内业空中三角测量技术的应用[J].铁道工程学报,2006,(1):101-106.

[56]王佩军.摄影测量学(测绘工程专业)[M].武汉大学出版社,2005.

[57]杨文环.结合空三与SFM的近景摄影测量点云获取研究[D].中国矿业大学.2016.

[58]曹宁.以DEM为高程辅助的弱交会卫星影像的区域网平差[D].兰州交通大学.2014.

[59]张梦花.基于无人机低空摄影测量成图实例研究[D].西安科技大学.2017.

[60]邓琴.基于D P Modeler的倾斜影像三维建模[D].东华理工大学.2015.

[61]卢晓攀.无人机低空摄影测量成图精度实证研究[D].中国矿业大学.2014.

[62]代成.基于无人机倾斜摄影的交通基础设施规划方法研究 ——以农村公路为例[D].北京交通大学.2019.