高级检索

基于宽基线图像远距离场景的自动三维重建

Automatic Reconstruction of Distant Scenes from Wide Baseline Images

  • 摘要: 图像获取时距离场景较远,由于图像对应点间视差较小,如果相机模型选取不合适,将导致本来不在同一个平面上的点重建在一个平面上.针对该问题,提出一种基于宽基线图像远距离场景的自动三维重建方法.该方法对相机模型的成像过程进行分析,给出了适用于远距离场景的透视相机模型;通过调整约束方程的权重系数,提高了仿射相机模型自标定过程的鲁棒性;利用分块的捆绑调整技术与分解法,解决了多视角重建结果融合的问题.室外远距离场景的重建实验结果表明,采用文中方法可取得较好的视觉效果.

     

    Abstract: When images are taken from a relatively long distance,the resultant disparity variation is usually small,consequently non-coplanar space points could be mistakenly reconstructed to a plane if the traditional pinhole camera model is employed,which will drastically affect the reconstructed scene visual appearance.Aimed at this problem,an automatic method for distant scene reconstruction is proposed in this work.The main contributions include:a more suitable perspective camera model for distant scene reconstruction,a robust self-calibration method based on affine camera model and the efficient fusion of reconstructed partial results is carried out via a combination of the factorization method and the bloc-based bundle adjustment technique.Extensive experiments with distant outdoor images are reported and validate our method.

     

/

返回文章
返回