基于随机抽取一致性的稳健点云平面拟合-北京工业大学学报.PDF

第40卷 第3期 北 京 工 业 大 学 学 报 Vol.40 No.3 2014年 3月 JOURNAL OF BEIJING UNIVERSITY OFTECHNOLOGY Mar. 2014 基于随机抽取一致性的稳健点云平面拟合 1,2 1 魏英姿 ,刘晓莉 (1郾 沈阳理工大学 信息科学与工程学院,沈阳摇 110159; 2郾 中国科学院沈阳自动化研究所 机器人学国家重点实验室,沈阳摇 110016) 摘摇 要:针对常用的平面拟合方法在点云数据存在误差或异常值时产生拟合不稳定的现象,提出了结合最小二乘 法的随机抽取一致性(random sample consensus,RANSAC)平面拟合算法. 该方法先用RANSAC算法检测并剔除异 常数据点,再利用最小二乘法将得到的有效数据点拟合,计算平面模型参数. 实验中,分别采用该算法和最小二乘 法、特征值法对仿真数据进行平面拟合,且采用本文提出的算法,分别对含有不同程度误差和异常值的点云数据进 行拟合计算. 研究结果表明:该算法适用于存在误差和异常值的点云数据拟合,能稳定地得到较好的平面参数估 值,具有较强的稳健性. 关键词:点云数据;随机抽取一致性(RANSAC);最小二乘法;平面拟合 中图分类号:TP301郾6 文献标志码:A 文章编号:0254-0037(2014)03-0400-04 Robust Plane Fitting of Point Clouds Based on RANSAC 1,2 1 WEI Ying鄄zi ,LIU Xiao鄄li (1郾School of Information Science and Engineering,Shenyang Ligong University,Shenyang 110159,China; 2郾State Key Laboratory of Robotics,Shenyang Institute of Automation,Chinese Academy of Science,Shenyang,110016,China) Abstract:In common plane fitting methods for point clouds,the results of planar parameter estimation are not always accurate when the gross errors and outliers are included. To overcome this shortcoming, RANSAC (random sample consensus) algorithm is proposed combined with least square method. The RANSAC algorithm is adopted to detect and eliminate the outlier points,and least square method makes plane fittingfortheremaininginnervalidpoints. Analyticalsimulationexperimentshavebeenconducted. Comparative results between our method and traditional methods,such as least squares and eigenvalue method,are provided. The proposed met