基于双目立体视觉的三维拼接和融合方法

3D registration and fusion based on binocular stereo system

为了获得更为全面的三维信息,本文提出了基于双目立体视觉图像的三维拼接和融合的解决方法,并且完成了算法的实现。本方法仅使用双目图像作为输入,完成了高效的空间转换,并侧重于利用重投影融合和场景信息建模的方法对空间开销的控制,提高了处理的效率。同时,为了方便三维信息的观察和效果显示,对三维数据进行了可视化。经实验证明,本方法不仅能够用于拼接和融合三维数据,更能较好地用于显示三维数据。

For getting more comprehensive 3D information , this paper presents a registration and fusion method based on binocular stereo vision and completes algorithm implementation This method uses only binocular image as input and completes efficient space transformation , focused on the use reprojection fusion and scene information modeling to control space cost as well as improving efficiency. Meanwhile , the method visualizes 3D data for the convenience of observation. The experiments show that this method can be used for 3D registration and fusion as well as display 3D data.