一种基于立体像对稠密匹配的三维重建方法

为了提高三维重建的性能,提出了一种新的基于立体像对稠密匹配的三维重建方法。该方法首先采用SIFT算法提取立体像对的特征点,并通过特征点的相似性判定进行特征匹配,然后从中选取种子点,并对种子点进行区域传播稠密匹配,再用对称对极点距离法消除误匹配得到精确匹配点对,然后利用摄像机矩阵计算出对应点的空间三维坐标,最后通过三角剖分和纹理渲染重建出三维模型。真实图像实验结果表明,这种重建方法显著优化了现有的基于稠密匹配算法的三维重建,且其效率和精度远远优于传统的基于稀疏匹配的重建方法。

3-dimensional reconstruction based on image dense stereo matching

A new 3-demensional （3D） reconstruction method based on image dense stereo matching is proposed in this paper to improve the performance of 3D reconstruction. Firstly, the feature-points of ster- eo images are extracted by SIFT algorithm and they are matched through similarity constraints. The second step is to select seed points to determine the performance of the algorithm and then we can do dense matc- hing with these seed points. Afterward, the false matches can be eliminated by symmetrical epipolar dis- tance algorithm and the 3D points coordinates can be calculated in virtue of camera matrix. Finally, the 3D model can be established by Delauny triangulation and texture mapping. Experiments with real images show that the method can optimize existing 3D reconstruction method based on dense stereo matching. Mo- reover, both the efficiency and precision of the method are better than those of traditional method based on sparse matching.