【作者】 李力；

【导师】 熊盛武； 段鹏飞；

【作者基本信息】 武汉理工大学 ， 计算机应用技术， 2014， 硕士

【摘要】 随着城市化进程的加速，通信网络和智能设备的发展以及移动互联网的崛起，人们的移动能力变得更强，对导航信息技术的需求也在日益增加。目前的导航应用大多只针对车辆，少数针对行人的导航软件也只是局限于室外，而没有将室内外结合起来，没有体现出行人和车辆导航的差异，导航体验较差。此外，室内条件下的定位对精度要求更高，目前使用较多的室内定位技术是基于WiFi的定位方法，它的精度依赖于场景中布置的大量AP以及后台提供的定位技术服务，一旦场景中AP热点不足或者服务中断，将无法获取在室内准确的位置，因此有必要研究新的能够辅助定位的方法。本文针对行人增强导航中出现的多种问题，结合图像匹配和增强引导相关的技术研究了如下内容：（1）提出了一种室内外一体的数据建模方法。依据采集的建筑物GIS数据以及场景中的图像数据，针对实验场景分别对室内和室外路网进行了数据建模，同时提出了结合增强要素的室内外数据组织和访问方法。（2）提出了使用图像匹配技术的室内外位置校正方法。将图像匹配技术应用于行人位置的校正，在数据存储上采用适合大数据并且读取高效的HDF5数据库，并建立特征库减少干扰信息，以提高匹配的准确率，同时降低匹配的计算量。（3）提出了导航中结合不同增强载体的增强引导方式。研究了使用实景和全景图两种载体的增强引导的关键技术，并依据导航中行人关注的要素提出了增强要素的实现方法和其同增强载体的融合方法。（4）实现了室内外一体的行人导航原型系统。本文以四维图新公司的办公楼及附近区域作为实验场景，在路径规划中提出了考虑道路重要性和建筑物出入口重要性的路径制定方法，实现了增强引导技术的应用，验证了本文提出的数据建模方法以及相关技术的可行性。更多还原

【Abstract】 With the development of urbanization, communication networks, smart devices,and the rise of the mobile Internet, the mobility of human being has become stronger,which resulted in the increasing demand for navigation information. Currently, mostof the navigation applications are used for vehicles, and only a few are designed forpedestrians which are limited to outdoors. Users have poor experience with theseapplications because indoor and outdoor information isn’t combined. What’s more,the difference between vehicle navigation and pedestrian navigation isn’t reflected inthose applications. Additionally, indoor localization has greater requirements inprecision. At present most indoor positioning technologies are based on Wi-Fipositioning methods, the accuracy depends on the arrangement of a large amount ofAccess Points (AP) in the scene as well as the positioning service running in thebackground. Once the AP was not sufficient or the positioning service was interrupted,it will be impossible to get accurate indoor positions. So it’s necessary to study newpositioning methods.The major work of this thesis is as follows:(1) According to the GIS data of the buildings and the corresponding image datain the scene, this thesis built the data model of indoor and outdoor road networkrespectively, and proposed a method to combine and enhance the indoor and outdoordata.(2) A correction method for both indoor and outdoor localization was proposedbased on the image matching technology. With this technology, the locations of thepedestrians could be corrected. And HDF5database was used because it worked wellfor huge data and had high efficiency for I/O operations. At the same time, a featuredatabase was established to reduce the impact of interference information, minimizecalculation and improve accuracy in the process of image matching.(3) This thesis studied the key technologies of augmented guidance and proposedthe implementation and integration methods with augmented carrier according to theelements of pedestrian concerned in the navigation.(4) The office building of NavInfo company and the surrounding area wereselected as the experimental scenario in this thesis. And a path planning algorithm which took the importance of roads as well as building entrances into considerationwas proposed. Finally a prototype navigation system for pedestrians was developedby combining both indoor and outdoor roadwork, which realized the enhancementguidance technology and verified the accuracy of the proposed data modeling methodas well as the feasibility of the corresponding technologies.更多还原