基于ASIFT图像匹配算法的三维重建  

Research on Technology of 3-D Reconstruction Based on ASIFT Image Matching Algorithm

作　　者：何永洁[1] 刘德宽[1] 陈孝威[1] 

机构地区：[1]贵州大学计算机科学与信息学院,贵州贵阳550025

出　　处：《贵州大学学报（自然科学版）》2012年第5期72-75,89,共5页Journal of Guizhou University:Natural Sciences

基　　金：贵州省优秀教育科技人才省长专项资金项目(200714)

摘　　要：提出了基于ASIFT图像匹配算法的三维重建算法。目前,基于图像序列的三维重建中,一般采用SIFT图像匹配算法。对于存在仿射变换的图像序列,ASIFT算法较SIFT算法能够获得更多精确的稀疏匹配点;基于ASIFT算法恢复的三维点云比基于SIFT算法恢复的三维点云更加稠密,从而能获得更好的三维重建结果。仿真实验表明,本文算法能获得较好的三维模型。The algorithm of 3-D reconstruction based on ASIFT image matching algorithm was put forward in this paper. At present, SIFT image matching algorithm is used generally in 3-D reconstruction method based on image sequence. For image sequence existing affine transform, ASIFT image matching algorithm can achieve more quantity and more accurate matches than SIFT image matching algorithm. The recovered 3-D point cloud based on ASIFT image matching algorithm is denser than that based on SIFT image matching algorithm, thus we can get a better 3-D reconstruction results. The simulation experiments show that a better 3-D model can be obtained by our method.

关 键 词：SIFT ASIFT 图像匹配 三维重建 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]