一种面向倾斜影像三维建模的影像筛选方法

摘要

倾斜摄影为城市三维建模提供更好数据源,同时增加摄影测量处理的数据量.文中提出一种基于自动空三结果稀疏点构建粗略表面模型的影像筛选方法,在保证精度相当情况下,减少参与三维重建的影像数据量,提高建模的效率.利用SWDC-5获取的航空倾斜摄影影像进行实验,当使用原始24.5% 的影像数据时,模型优化时间减少为原来的36.1%,除个别植被区域外,利用筛选后影像重建的模型精度与采用全部影像重建的结果相当.%Oblique photogrammetry provides a better data source for urban 3D modeling,however it increases the amount of image data for photogrammetry processing.In this paper,an image selection method based on sparse point cloud from automatic aerial triangulation is proposed.T he method can reduce the amount of image data involved in 3D reconstruction and improve the efficiency of modeling. When only 24.5% of original images is used,the time of model optimization is reduced to 36.1%.Apart from some vegetation areas,the accuracy of the model is compared with the model reconstructed with all images.