【作者】 高扬；

【导师】 杨志强；

【作者基本信息】 长安大学 ， 大地测量学与测量工程， 2013， 博士

【摘要】 卫星导航和位置服务融合了信息业、制造业、服务业等多行业，也成为智能手机等终端设备的必备功能。随着移动互联网生态圈的形成，卫星导航定位和基于位置的服务(LBS)必将给互联网、无线通信、物联网、车联网带来日新月异的变化，改变日常生活方式。作为国家大力培育的战略性新兴产业，卫星导航与位置服务也是世界各国抢占未来竞争的战略制高点，正成为继互联网、移动通信之后的全球第三大信息技术和新的增长点。快速发展和应用的移动互联网络时代，用户体验直接决定一个产品或应用模式的生存。在导航定位与位置服务中，用户需要在路径引导过程中采用更加接近人们空间交际的方式组织和表达路径信息，以直观、渐进、多模式的表达多源导航信息，提高导航认知的可视性、真实性和可靠性，提高导航位置服务用户体验。可量测实景影像综合了海量实景影像可视化和可量测定位的特点，为位置导航提供了一种全新数据源，一方面可以提高位置导航应用的体验和真实性，为导航提供了更多、更高分辨率和更精细的数据资源，另一方面，可量测实景影像具有严密的内、外方位元素，可基于可量测实景影像进行测量计算，通过可量测实景影像与实时影像、卫星定位数据和惯导数据的融合处理，为提高导航定位的精准性和可靠性奠定了基础。本文着眼于来自车载移动道路测量的海量可量测影像在导航定位和实景位置导航可视化服务两方面的技术开展研究，将可量测影像与二维导航数字地图、卫星导航定位数据、CCD实时影像数据、惯导数据和里程数据有机融合，为实景导航位置服务提供一种基于可量测影像实景导航的新思路。本文主要开展了如下研究工作：1）从空间认知入手，系统探讨了各种位置导航方法，提出了可量测影像实景导航概念和特点，分析了基于可量测影像的实景位置导航的必要性和可行性。研究了基于车载移动测量系统的可量测影像采集与更新技术，在此基础上提出了基于可量测影像的多源信息融合实景导航技术框架。2）研究了基于可量测影像的多源信息融合实景导航数据模型。针对可量测影像实景导航数据的特点，提出了实景导航地图的概念模型，进一步研究了实景导航地图数据组织和高效检索显示方法。提出了二维导航数字地图的“多级四叉树+R树+LOD”的索引结构、可量测实景影像的动态分段索引结构和数据分层存储模型，实现了可量测实景影像与二维导航数字地图的有效融合，提高了数据检索效率，适应了可量测实景影像的多种应用模式。在此基础上研究了车道级路线导引与可量测实景影像相集成的实景导航表达方法。3）在可量测影像实景导航定位方面，研究了基于可量测实景影像与实时影像匹配的导航定位方法。由于实时影像与可量测实景影像获取时具有不同的光照、视角，给影像匹配带来困难。本文提出了一种旋转尺度不变的SIFT特征进行两种影像的匹配，有效解决不同照度、不同视角下的两种影像匹配的关键问题。进一步提出了基于可量测实景影像匹配的车辆导航定位方法。通过可量测实景影像与实时影像匹配实验及定位实验验证了方法的合理性和有效性。4）研究了可量测实景影像实时匹配定位结果与卫星定位/惯性导航/里程计等多源数据融合处理导航定位方法。在对IMU零偏温度特性分析基础上，提出了IMU零温补偿模型，并通过实验验证了模型的可靠性。进一步研究了可量测实景影像实时匹配定位与GPS、IMU和里程计数据的联邦Kalman滤波导航计算方法，通过实际跑车实验验证了方法的有效性。5）可量测影像实景导航原型构建与实验验证工作：以基于可量测实景影像的某型武器无依托发射导航系统原型为例，对可量测影像实景导航的技术体系、可量测影像实景导航的数据组织、实时匹配定位和融合处理导航计算等关键技术进行了试验与精度分析，验证本文提出的可量测影像实景导航方法的可行性和关键技术的有效性。更多还原

【Abstract】 Satellite navigation and location based services are integrated informationtechinique, manufacturing and services, which have become the basic function of smartphones. With the rapid development of mobile internet, satellite navigation andlocation-based services (LBS) will bring new changes of Internet, wirelesscommunications, the Internet of things and Telematics in daily life style. Satellitenavigation and location based services is becoming the new cutting edge technique afterinternet technique and wireless communication technique in the IT field worldwide. Inthe rapid developing mobile Internet era, the user experience of a product or applicationdirectly determines its survival mode. For the navigation and location based services,the users need more guidance information and high precision position to improve thevisualization, authenticity and reliability of location-based services to obtain high userexperience. Digital measurement image (DMI) combine image visualization andmeasurement position feature together, which become a new data for navigation andLBS. The digital images can improve experience and authenticity of navigation. On theother hand, the DMIs have direct geo-referenced information, thus we can use DMIs formeasurement and they can integrate with real time images, satellite position data andinertial measurement unit data to improve the accuracy and reliability of navigation andpositioning system.In this thesis, a digital measurement images from land based mobile mappingsystem are introduced into navigation both for navigation visualization and position.The DMIs,2D navigation map, satellite position data, real time images from CCDcamera and inertial measurement unit data are integrated together for a new imagenavigation approach in the following aspects:1) The current navigation technique and new requirements are discussed from thespatial cognition. A new DMIs based image navigation concept is proposed. Thenecessity and feasibility of image navigation based DMIs are analyzed. The DMIsacquisition and updating technology based on land based mobile mapping system arediscussed.Anew image navigation technique framework is presented. 2) A digital measuable image navigation data model is proposed. The DMIs datamodel together with high efficient index method is presented. A two dimensionnavigation data model based on multiple level quadtree, R tree and LOD (level of detail)is proposed to obtain high efficiency display. A DMIs index approach based on roaddynamic segment is presented for DMI query. And a layer pyramid model based onlarge file is designed for DMIs storage. A route guidance method on the DMIs isdiscussed to achieve image based navigation visualization.3) A navigation position approach based on real time image and DMIs ret rivalingand matching is proposed in this thesis. Considering the image matching difficulty withvarious illumination and perspective, a SIFT (Scale-invariant feature transform) imagematching algorithm is presented for real time image matching with DMIs. A navigationopposition approach based on real time image and DMIs matching is proposed. Anexperiment is designed to verify the rationality and effectiveness of the proposedmethod.4) A new navigation position approach based on DMIs matching position, satelliteposition data, inertial measurement position data and odometer data fusion is presentedin this thesis. An IMU (inertial measurement unit) zero bias model is proposed based onthe IMU zero temperature characteristic analyst. An experiment is designed to verify thereliability of the proposed IMU zero bias model. A Federated Kalman Filter methodwith DMIs image position result, GPS, IMU and odometer is designed to achieve DMIsbased position. And the experiment result shows that the DMIs aided position canimprove the navigation position accuracy and reliability.5) An image navigation prototype for a certain type of weapon without relying isdesigned and presented for the proposed technique experiment and verify in this thesis.The experiment result show that the proposed DMIs based image navigation canachieve high position accuracy, reliable navigation and true3D visualization.更多还原