基于RANSAC的奇异值剔除的单目视觉里程计  

Monocular visual odometry with RANSAC-based outlier rejection

作　　者：孙作雷[1] 黄嘉明[1] 张波[2] SUN Zuolei HUANG Jiaming ZHANG Bo(Information Engineering College, Shanghai Maritime University, Shanghai 201306, China Shanghai Advanced Research Institute, Chinese Academy of Sciences, Shanghai 201210, China)

机构地区：[1]上海海事大学信息工程学院,上海201306 [2]中国科学院上海高等研究院,上海201210

出　　处：《上海海事大学学报》2016年第4期87-91,共5页Journal of Shanghai Maritime University

基　　金：国家自然科学基金(61105097;51279098;61401270);上海市教育委员会科研创新项目(13YZ081)

摘　　要：为提高单目视觉里程计算法的性能,从视觉特征选取和特征误匹配剔除两个方面进行研究.采用SURF描述子提取单目图像的特征点,并匹配相邻图像序列的特征,使用归一化线性八点法依次得到基础矩阵和本质矩阵.利用三角测量求解匹配点的三维坐标,进而根据2D-2D模型解算出两帧图像间相机运动的旋转和平移,从而构建单目视觉里程计系统.为提高算法性能,使用RANSAC算法清除初次计算的特征误匹配,并利用地面数据获取相机运动的平移尺度.实验结果验证了RANSAC算法能够有效剔除特征误匹配,降低单目视觉里程计的累积误差.In order to enhance the algorithm performance of the monocular visual odometry, the visual feature extraction and the mismatched feature rejection are studied. The SURF descriptor is employed to extract features of monocular images and match features in the adjacent image sequence. The fundamental matrix and essential matrix are derived using the normalized eight-point method. The 3D coordinates of matching points are calculated with the triangulation, and then the camera translation and rotation between two frames of images are estimated based on 2D-2D model. As a result, the system of monocular visual odometry is constructed. To improve the algorithm performance, RANSAC algorithm is adopted to reject the feature mismatching in the first calculation, and the camera translation scale is achieved by the ground data. The experiment result demonstrates that RANSAC algorithm can effectively eliminate feature mismatching and reduce the cumulative error of the monocular visual odometry.

关 键 词：机器人定位 视觉里程计 特征提纯 机器视觉 SURF RANSAC 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]