基于全信息摄影测量方法的复杂建筑三维重建

Three-Dimensional Reconstruction of Complex Construction Based on Full Information Photogrammetry

摘要: 针对被摄对象受遮挡导致的影像信息不全、不同数据集的影像融合空三易失败等问题，本文给出了一种全信息影像采集与数据处理方法。该方法对影像采集和多源影像空三解算的关键技术进行优化，给出了空中多视倾斜影像加地面多层多角度的全信息影像采集方法和分步整体联合空三的平差解算方法。以某复杂建筑为研究对象，利用Phantom 4 RTK无人机采集全信息影像数据，综合使用ContextCapture和MicMac软件及编程语言进行实景三维重建。结果表明：与传统建模方法相比，该方法在位置精度、点云偏差、白膜结构和纹理细节几个方面都表现优越，解决了复杂建筑模型数字化、真实再现等关键难题，为智慧水利复杂建筑物建模提供支持。

Abstract: In response to issues such as incomplete image information due to subject obscuration, and the aerial triangulation failure of merging imagery from different datasets, this paper proposes a comprehensive method for image acquisition and data processing. This method optimizes key techniques in image acquisition and multi-source image aerial triangulation, and gives the full information image acquisition method of aerial multi-view oblique image plus ground-based multi-layer, multi-angle close-range image, adjustment solution method of stepwise integrated aerial triangulation. To validate the effectiveness of this method, this paper conducted a study focusing on a complex ancient building. The full-information image data was collected using a Phantom 4 RTK drone, and a combination of ContextCapture, MicMac software, and programming languages were employed for real-time three-dimensional reconstruction. The results demonstrate the superiority of the method compared to traditional modeling approaches in terms of positional accuracy, point cloud deviation, white film structure, and preserving textural nuances. This method effectively addresses key challenges in digitizing and realistically reproducing complex construction, making it highly significant for modelling complex construction in smart water resources.