倾斜摄影与地面移动测量组网观测数据融合处理技术研究

Aiming at the necessity of obtaining the three - dimensional spatial information by the combination of oblique photography and vehicle-borne mobile mapping, a new idea of the observation of the integrated spatial-terrestrial networks is put forward, which makes up the limitations of these two kinds of technologies. Aiming for handling the problem of large change in field angle, large scale change and large change in illumination, this paper studies the technology in fusion of two kinds of data. Focusing on breakthroughs in the feature matching technology in images with large field angle changes in super wide baseline imaging , "moving - moving" combination of high-precision baseline constraints of the network observation of the overall aerial triangulation adjustment theory, dense matching technology in images observed by integrated spatial-terrestrial networks and image quality improvement technology ,this paper proposed a system of overall calculation of data in integrated spatial-terrestrial networks instead of the current independent processing mode of processing before integration in oblique photography and vehicle-borne mobile mapping. This system is an effective solution to problem of low quality and low efficiency in three-dimensional reconstruction of urban areas due to incomplete data sources. The integrated spatial-terrestrial networks can promote the rapid development of technology in obtaining urban spatial information and the integrated application of the oblique photography and vehicle-borne mobile mapping to improve the quality and efficiency of urban 3D reconstruction. We will publish 3 to 5 SCI papers, 3 to 5 papers of core journals in China, 1 invention patent, 1 monograph and cultivate 3 to 6 postgraduates.

针对倾斜摄影与车载移动测量联合获取三维空间信息的必要性，提出了空地组网观测的新思路，弥补了这两类技术各自存在的局限性。针对空地组网观测成像视角变化大、尺度变化大和光照变化大带来的整体处理技术难题，开展两类数据融合处理共性技术研究。重点突破超宽基线成像大视角变化的影像特征匹配技术、“动-动”组合高精度基线约束的组网观测整体空三解算理论、空地组网观测影像密集匹配技术以及影像质量改善技术等。提出的空地组网观测数据整体解算技术体系，替代了目前倾斜摄影与车载移动测量的“先处理、后整合”的独立处理模式，有效解决数据源不完备带来的城市三维重建低质量、低效率难题。空地组网能够促进城市空间信息快速获取技术的发展以及倾斜摄影与车载移动测量技术的综合应用，提高城市三维重建生产的质量和效率。发表SCI论文3～5篇，国内核心期刊论文3～5篇，申请1项发明专利，出版专著1本，培养研究生3～6名。

倾斜摄影测量和地面车载移动测量是进行城市三维建模的两种重要技术手段，项目针对单独使用倾斜摄影测量技术以及地面车载移动测量技术进行城市三维建模中存在的数据不完备等局限性，提出进行空地组网观测、融合倾斜摄影测量数据与地面车载移动测量数据，实现城市的完备三维建模，提高城市三维建模的质量和效率。.倾斜摄影测量数据与地面车载移动测量数据的联合处理是实现空地联合建模的关键，由于倾斜摄影测量数据与地面车载移动测量数据的获取视角不同、影像尺度不一致，导致倾斜摄影测量数据与地面车载移动测量数据联合处理困难。为实现空地数据融合、空地数据联合建模，项目研究了空地组网模式及其控制技术、空地组网影像特征匹配技术、空地组网整体空三解算理论、空地组网密集匹配技术、空地组网影像质量改善技术等内容。.经过四年的研究，项目集成了倾斜摄影测量系统和地面车载移动测量系统，实现了“1+1”的空地组网控制与数据获取；提出了基于ASIFT的空地组网影像特征匹配方法以及多层次空地组网连接点匹配方法，实现了空地组网影像的稳健连接点匹配；研究了带附加参数的自检校区域网平差方法，将自检校参数、控制点坐标及外方位元素视为带权观测值，采用附加约束参数模型，实现了空地组网影像的联合空三；提出了基于结构相似性测度的空地影像密集匹配技术以及基于PatchMatch顾及多尺度几何约束的弱纹理区域密集匹配方法，实现了联合空地组网影像的三维重建。研发了倾斜影像与地面移动测量组网观测数据融合处理软件MoGAS。.项目研究成果发表期刊论文16篇（其中SCI论文7篇）、会议论文2篇，资助6位博士、9位硕士完成学位论文，资助8人次赴国外参加学术会议。