一种改进SIFT算法的单目视觉里程计-《集美大学学报（自然科学版）》

文章信息/Info

Author(s):: ZHAO Li-ming; CHEN Ning; (School of Mechanical and Energy Engineering,Jimei University,Xiamen 361021,China)

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，提出一种改进SIFT特征点匹配的单目车载视觉里程计算法.首先，为了提高特征点匹配的正确率和实时性，结合主成分分析法和平面极线几何约束，改进了传统SIFI匹配算法，其次，建立合理的移动机器人运动数学模型，得到连续帧间图像信息和移动机器人运动位姿变化的转换关系.试验结果表明，误差仅为1.6%，算法运行时间缩短0.022 s.

Abstract:: This paper presents a robust method for monocular visual odometry capable of accurate position estimation of a mobile robot based on an improved SIFT algorithm.Firstly,for improving the accuracy and performance of real-time of feature points matching,improvements have been made on SIFT algorithm by combining Principal Component Analysis and Epipolar geometric constraint.Secondly,a mathematical model has been developed to determine the relationship between the information of consecutive frames and pose change of mobile robot.Results show that the algorithm brought about an accuracy with relative error less than 1.6%,and the running time of the algorithm reduces 0.022 seconds correspondingly.