基于标识的双目视觉室内定位算法研究

发布时间：2018-04-08 19:29

本文选题：双目视觉室内定位　切入点：标识　出处：《哈尔滨工业大学》2017年硕士论文

【摘要】：互联网和可穿戴设备的飞速发展,使得位置感知逐渐成为人们必不可少的信息,定位服务正逐渐得到研究人员的广泛关注。由于人们70%以上的时间处于室内环境,因此室内定位技术成为人们目前的研究热点。随着手机、相机等智能终端的普及,其完备的内置传感器设备使得基于视觉的定位系统正逐渐替代那些需要特殊设计的室内定位系统。本文首先研究了双目视觉定位技术的目的与意义,分析了视觉定位技术以及双目视觉定位技术的国内外研究现状。其次,分析了双目视觉系统模型和图像局部特征等相关理论基础在视觉定位系统的应用。除此之外,本文针对传统视觉定位系统在离线阶段和在线阶段存在的一些问题完成了如下研究:(1)针对传统位置指纹离线数据库建立过程中工作量过大的问题,提出了一种基于标识的离线数据库建立方法,该方法利用场景图像中标识区域的SURF局部特征点代替全局位置指纹信息,能够有效地减少数据库的存储容量,并通过图像间的单应性映射减少了建立阶段的复杂度,节省了时间开销;(2)针对传统基于标识的单目视觉定位算法中定位精度低的问题,提出了一种改进的双目视觉定位算法,该算法有效地减少了标识特征匹配的时间开销,提高了定位精度,同时该算法还能够计算出用户转向角,丰富了定位导航信息。对提出的基于标识的离线数据库的数据库容量和一种改进的双目视觉定位算法的定位精度进行实验仿真。仿真结果表明本文提出的离线数据库建立方法能够有效地减少数据库的存储容量,减少了建立阶段的时间复杂度,节省人力开销;改进的双目视觉定位算法可以提高在线用户的定位精度。
[Abstract]:With the rapid development of the Internet and wearable devices, location perception has gradually become an essential information for people, and location services are being paid more and more attention by researchers.Because people spend more than 70% of their time in indoor environment, indoor positioning technology has become a hot research topic.With the popularity of mobile phones, cameras and other intelligent terminals, its complete built-in sensor equipment makes the vision-based positioning system is gradually replacing those indoor positioning systems that need special design.In this paper, the purpose and significance of binocular vision positioning technology are studied, and the present situation of vision location technology and binocular vision positioning technology at home and abroad is analyzed.Secondly, the application of binocular vision system model and image local feature in vision positioning system is analyzed.In addition, aiming at some problems existing in the off-line and on-line phase of traditional visual positioning system, this paper has completed the following research: (1) aiming at the problem of too much work in the process of setting up the traditional location fingerprint off-line database,In this paper, an off-line database building method based on identification is proposed. This method can effectively reduce the storage capacity of the database by replacing the global location fingerprint information with the SURF local feature points of the identification region in the scene image.Through the monogram mapping between images, the complexity of the establishment stage is reduced, and the time cost is saved. Aiming at the problem of low localization accuracy in the traditional monocular vision localization algorithm based on identification, an improved binocular vision location algorithm is proposed.The algorithm can effectively reduce the time cost of identification feature matching and improve the positioning accuracy. At the same time, the algorithm can calculate the user steering angle and enrich the location and navigation information.The database capacity of the proposed off-line database based on identification and the localization accuracy of an improved binocular vision location algorithm are simulated experimentally.The simulation results show that the off-line database establishment method proposed in this paper can effectively reduce the storage capacity of the database, reduce the time complexity of the establishment phase, and save the cost of manpower.The improved binocular vision location algorithm can improve the location accuracy of online users.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41

【参考文献】