基于Virtuoso3D的大面积弱纹理区域空三计算研究

Research on Aerial Triangulation Calculation of Large-area Weak Texture Region Based on Virtuoso3D

摘要:

针对目前无人机倾斜摄影测量内业数据处理过程中，大面积弱纹理区域空中三角测量出错率高且精度难以保证的现状，现采用Virtuoso3D对大面积弱纹理区域进行空中三角测量计算。首先，介绍了基于Virtuoso3D平台的空中三角测量的操作步骤，以容东片区实际项目为例，该区域被大面积植被、防尘网覆盖，航拍过程中会出现在多张连续影像中，存在特征点不明显或者缺失现象。在空三计算过程中，基于多视影像进行特征点计算提取的方法容易出现连续弱纹理像片不参与运算出现漏洞的问题，而Virtuoso3D采用机器视觉普适性算法，配合基于摄影测量的严密解算方式，在减少匹配次数的同时充分提升空三成功率和精度，有效解决大面积弱纹理区域空三连接点分布不均、断裂、分层等问题，自由网之后的精转点遵循由粗到精的原则分层迭代获得可靠的空三成果。

Abstract:

Aiming at the current unmanned aerial vehicle (UAV) tilt photogrammetry in-house data processing, the error rate of the aerial triangulation is high in the large-area weak texture region, and its precision is hard to guarantee the present situation. The aerial triangulation calculation is carried out for the large-area weak texture region by using Virtuoso3D. Firstly, the operating sequence of aerial triangulation based on Virtuoso3D is introduced. Taking an actual project in Rongdong Area as an example, there are the unapparent feature points or deficiency phenomenon in multiple continuous images during the aerial photo because of this region covered by large-area vegetation and dust screen. In the process of aerial triangulation calculation, the method of calculating and extracting the feature points based on multi-view images is easy to appear the problem that the continuous weak texture image does not participate in the operation. And Virtuoso3D adopts the universality algorithm of machine vision to match the rigorous solution based on photogrammetry. The success rate and accuracy of aerial triangulation calculation are fully improved at the same time of reducing the matching times, which effectively solve the problems of uneven distribution, fracture and layering of aerial triangulation connection points in large-areas weak texture region. The fine turning point after the free net follows the principle of layering and iterating from coarse to fine to obtain the reliable aerial triangulation results.