| 摘 要: 在视觉SLAM前端特征点匹配过程中，采用RANSAC算法剔除误匹配特征点存在迭代次数不稳定、 效率低、鲁棒相差等问题，从而对相机定位产生影响。与ORB算法结合，本文引入一种渐进采样一致性算法，即 PROSAC(Progressive Sampling Consensus)，来消除迭代次数不稳定问题。利用Kinect v2相机对改进的RGB-D SLAM算法进行实验，获得三维点云地图和相机轨迹，实现了ORB+PROSAC的误匹配剔除算法。与ORB+RANSAC的 结合方式相对比，本文算法验证鲁棒性更好，实时性更强。 |
| 关键词: 视觉SLAM；特征点匹配；RANSAC算法；PROSAC算法 |
|
中图分类号: TP391.4
文献标识码: A
|
|
| A Study of Visual SLAM Based on ORB+PROSAC Mismatch Elimination Algorithm |
|
XU Zifeng,SHI Chao,WANG Yongfeng,CHEN Long
|
( College of Mechanical Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China)
|
| Abstract: In the process of feature point matching in the front of visual SLAM,a series of problems have been found when using the RANSAC algorithm to eliminate the mismatched feature points,such as unstable iterations,low efficiency and robust phase difference,resulting in an impact on camera positioning.Combined with the ORB algorithm,this paper introduces a progressive sampling consensus algorithm,PROSAC (Progressive Sampling Consensus),to eliminate the instability of iteration times.Using the Kinect v2 camera with the improved RGB-D SLAM algorithm,the 3D point cloud map and camera trajectory can be obtained in the experiment and the mismatch elimination algorithm of ORB+PROSAC can be realized. Compared with the combination with ORB+RANSAC,the proposed algorithm verifies the robustness better with a stronger real-time performance. |
| Keywords: visual SLAM;feature point matching;RANSAC algorithm;PROSAC algorithm |
