一种基于局部特征的栅格地图拼接方法  

A GRID MAP MERGING APPROACH BASED ON LOCAL FEATURE

作　　者：李雅梅[1] 苏龙 Li Yamei;Su Long(Faculty of Electrical and Control Engineering,Liaoning Technical University,Huludao 125105,Liaoning,China)

机构地区：[1]辽宁工程技术大学电气与控制工程学院

出　　处：《计算机应用与软件》2020年第1期110-115,共6页Computer Applications and Software

摘　　要：栅格地图拼接是多机器人协作创建地图的必不可少的环节。提出一种基于局部特征的栅格地图拼接方法,旨在克服传统算法拼接效率低、拼接成功率低等问题。该方法先对栅格地图拼接建立数学模型,然后转化为图像匹配问题。搭建非线性金字塔,利用FAST算法定位特征点,使用PCA-SIFT算法建立描述符进行匹配,借助随机采样一致性优选匹配点,得到初始参数,并提出一种新型的栅格地图拼接规则。实验表明,该方法鲁棒性好、拼接速度快、拼接精度高。Grid map merging is an essential part in cooperative mapping for multi-root systems.We study a grid map merging method based on local features so as to overcome the problems of low merging efficiency and low merging success rate of traditional algorithms.We established the mathematical model for the grid map merging,and translated it into an image matching problem.The nonlinear pyramid was built,and FAST algorithm was used to locate the feature points.We adopted PCA-SIFT algorithm to set up descriptors for matching,employed RANSAC(random sampling consensus)algorithm to optimize matching points and get initial merging parameters,and proposed a new rule for grip map merging.Experiments show that the method has good robustness,fast merging speed and high merging accuracy.

关 键 词：栅格地图拼接 图像匹配 非线性金字塔 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]