计算机工程与应用 ›› 2011, Vol. 47 ›› Issue (14): 170-173.

• 图形、图像、模式识别 • 上一篇    下一篇

基于改进SIFT算法的双目视觉SLAM研究

朱代先1,王晓华2   

  1. 1.西安科技大学 通信与信息工程学院,西安 710054
    2.西安工程大学 电信学院,西安 710048
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2011-05-11 发布日期:2011-05-11

Research of Binocular vision SLAM algorithm based on improved SIFT

ZHU Daixian1,WANG Xiaohua2   

  1. 1.Communication and Information Engineering College,Xidian University,Xi’an 710054,China
    2.College of Electronics and Information,Xi’an Polytechnic University,Xi’an 710048,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-05-11 Published:2011-05-11

摘要: SIFT算法通常用于移动机器人视觉S LAM中。但其算法复杂、计算时间长,影响视觉SLAM的性能。在两方面对SIFT改进:一是用街区距离与棋盘距离的线性组合作为相似性度量;二是采用部分特征方法完成快速匹配。应用扩展卡尔曼滤波器融合SIFT特征信息与机器人位姿信息完成SLAM。仿真实验表明,在未知室内环境下,该算法运行时间短,定位精度高。

关键词: 尺度不变特征变换(SLAM), 同步定位与地图构建(SIFT), 双目视觉, 扩展卡尔曼滤波

Abstract: Scale Invariant Feature Transform(SIFT) algorithm is used in mobile robot Simultaneous Localization and Mapping(SLAM) based on visual information.but this algorithm is complicated and computation time is long.Two improvements are introduced to optimize its performance.The linear combination of cityblock distance and chessboard distance is comparability measurement;Some partial features are used to matching.SLAM is completed by fusing the information of SIFT features and robot information with EKF.The simulation experiment indicate that the proposed method reduce computational complexity,and with high localization precision in indoor environments.

Key words: Simultaneous Localization and Mapping(SLAM), Scale Invariant Feature Transform(SIFT), binocular vision, extended Kalman filter