基于LM和SVD结合的点云拼接算法研究  

Research on the Point Cloud Stitching Algorithm Based on Combination of LM and SVD

作　　者：吴晓庆[1] 黄玉清[1] 

机构地区：[1]西南科技大学信息工程学院,四川绵阳621010

出　　处：《自动化仪表》2017年第1期34-37,共4页Process Automation Instrumentation

基　　金：国家自然科学基金资助项目(61401379);国防应用技术项目资助(12zg610303)

摘　　要：点云拼接是场景三维重建过程中的重要环节。针对现有的拼接算法存在耗时长、拼接效果差等问题,通过利用正交映射,获取三维点云与二维图像之间的映射关系,并在二维图像中获取匹配点对。在此基础上,提出了采用LM和SVD结合的方法来变换矩阵。首先根据正交映射,获取三维点云所对应的二维图像;其次利用SURF算法,提取图像的特征点,并通过FLANN算法以及RANSAC算法,提取两幅图像中的特征拼接点对;最后将LM和SVD两种方法相结合,求得变换矩阵。试验结果表明,无论拼接精度,还是运行速度,该算法均较直接利用LM算法或SVD算法有所提高。Point cloud registration is an important role in the reconstruction of 3D scenes. Aiming at the problems existing in current registration algorithms, such as time - consuming and poor registration effect, through adopting orthogonal mapping, the mapping relations between 3D point cloud and 2D image is obtained, and the matching point pairs are captured in 2D image. On this basis, by combining levenberg - marquardt （ LM） algorithm and singular value decomposition （ SVD） , the matrix is transformed. Firstly, in accordance with the orthogonal mapping, the 2D image corresponding to the 3D point cloud is acquired; then the image feature points are extracted by using SURF algorithm, and with FLANN algorithm and RANSAC algorithm, the splicing point pair for two images is extracted; finally, the transformation matrix is obtained by combining the LM and SVD. The results of the tests indicate that the method proposed gets better splicing accuracy and running speed then those by LM algorithm or SVD algorithm.

关 键 词：机器人 点云拼接 正交映射 线性最小二乘法 LM算法 SVD算法 变换矩阵 

分 类 号：TH-3[机械工程] TP391[自动化与计算机技术—计算机应用技术]