Abstract: This paper presents a robust method for monocular visual odometry capable of accurate position estimation of a mobile robot, and this algorithm uses a single camera to extract SURF（Speeded Up Robust Features） features points on the road of the consecutive frames. Matching the feature points of two consecutive frames, it combines with epipolar geometry, and then gets the pose transformation matrix of the vehicle based on homography matrix. The experimental results show that this algorithm has high accuracy and a good performance of real-time.

Key words: visual odometry, epipolar geometry, homography, Speeded Up Robust Features（SURF）

摘要： 针对如何准确获取位姿信息来实现移动机器人的避障问题，提出一种可用于实时获取移动机器人位姿的单目视觉里程计算法。该算法利用单目摄像机获取连续帧间图像路面SURF（Speeded Up Robust Features）特征点；并结合极线几何约束来解决路面特征点匹配较难的问题，通过计算平面单应性矩阵获取移动机器人的位姿变化。实验结果表明该算法具有较高的精度和实时性。

关键词: 视觉里程计, 极线几何, 平面单应性, 加速鲁棒特征（SURF）