摘要
为了提高三维重建中双目特征匹配的匹配效率和重建质量,在基于传统的加速鲁棒特征(SURF)匹配算法基础上,提出了一种基于SURF-RANSAC配准的三维重建算法。利用左右两幅图像来进行三维重建,首先通过Hessian矩阵来获取目标图像的初始特征点,并用邻近快速搜索算法完成初步的特征点匹配,然后融合随机抽样一致性算法(RANSAC)来优化匹配,最后利用三维坐标和纹理映射来完成三维重建。在Open CV上对该算法进行验证。结果表明,本文算法比传统的三维重建算法具有更高的精确度和更快的速度。
In order to improve the matching efficiency and reconstruction quality of binocular feature matching in 3 D reconstruction,a speed up robust features( SURF)-random sample consensus( RANSAC) based on the traditional SURF matching algorithm was proposed,which was used the left and right images for 3 D reconstruction.Firstly,the initial feature points of the target image were obtained by Hessian matrix,and the preliminary feature point matching was completed by the fast library for approximate nearest neighbors algorithm. Then,the RANSAC was merged to optimize the matching. Finally,3 D coordinates and texture mapping were used to complete the 3 D reconstruction. The algorithm was verified on Open CV. The results show that the proposed algorithm has higher accuracy and faster speed than the traditional 3 D reconstruction algorithm.
作者
别治峰
刘守山
黄春凤
BIE Zhi-feng;LIU Shou-shan;HUANG Chun-feng(College of Electronic Information Engineering,Shandong University of Science and Technology,Qingdao 266590,China)
出处
《科学技术与工程》
北大核心
2019年第28期239-244,共6页
Science Technology and Engineering
基金
山东省重点研发计划(2015GSF118094)资助
