基于序列图像的三维重构

提出了一种基于序列图像的三维重构方法。在相机标定的基础上,利用两幅图像间的匹配关系得到一个初始的三维模型,然后利用剩余图像与初始模型间的匹配关系,采用反投影映射的方法计算出剩余图像的外方位元素。最后经过稠密匹配、三角剖分和纹理渲染得到物体的三维模型。实验结果表明,相机标定精度较高的情况下,采用该算法能够得到较低失真的三维模型。

The Method for Image Sequence of 3D Reconstruction

This paper presents a 3D reconstruction algorithm based on sequential image.The initial reconstruction is performed using the matching relations between two images based on the camera calibration.Then the external parameters of the remained images are generated by means of reprojection and the matching relations between the initial reconstruction and the remained images.Finally,through dense matching、triangle meshing and texture rendering,the model of the object is obtained.Experimental results show that this method can reconstruct the 3D model with low distortion in the case of high precision of camera calibration.