移动增强现实跟踪注册技术的研究与应用

发布时间：2018-04-17 21:30

本文选题：移动增强现实 + 尺度空间　；参考：《沈阳工业大学》2017年硕士论文

【摘要】：增强现实是一种利用计算机产生的附加信息对使用者观察到的真实场景进行增强或扩展的一门综合性技术,是虚拟现实的衍生学科,也是近年来的一个研究热点。随着移动计算技术的发展,移动终端硬件性能的提高以及人们对智能手机的广泛使用,将增强现实系统应用于智能手机上成为了可能。移动增强现实是将增强现实技术有效地应用于移动设备终端,以移动设备为媒介实现虚拟与现实之间的交互。增强现实系统中,跟踪注册技术是增强现实系统构建的基础技术,也是决定增强现实系统性能的关键。传统的基于计算机视觉的跟踪注册算法计算复杂,耗时较长。与计算机相比,目前的智能手机内存较小、处理能力较低,不足以运行复杂度较高的算法。手机端增强现实系统的准确性和实时性都无法与PC端增强现实相比。本文因此针对移动增强现实技术的性能及功能需求,通过对各种特征点匹配算法进行对比实验,综合比较各种特征点匹配算法的优缺点,选择ORB算法作为本移动增强现实系统的特征点匹配算法。深入分析ORB算法,针对ORB算法检测出的特征点对尺度变化敏感的问题,提出结合尺度空间思想对ORB算法加以改进。在特征点匹配和位姿跟踪阶段,传统的ORB算法采用穷举法进行特征点匹配,所需计算量较大,不适用于手机端。因此引入视觉词袋模型的方法改变特征搜索策略,从而加快特征点匹配效率,计算摄像机位姿,实现对手机位姿的实时跟踪。设计实现了移动增强现实系统,提出“识别+跟踪”的移动增强现实系统框架,将改进后的算法应用于系统中。从而提高手机端算法的鲁棒性、实时性,使移动增强现实系统适用于大规模的室内外场景。
[Abstract]:Augmented reality (AR) is a kind of comprehensive technology which uses the additional information generated by computer to enhance or expand the real scene observed by users. It is a derivative subject of virtual reality and a research hotspot in recent years.With the development of mobile computing technology, the improvement of hardware performance of mobile terminal and the widespread use of smart phone, it is possible to apply augmented reality system to smart phone.Mobile augmented reality is an effective application of augmented reality technology to mobile device terminals, using mobile devices as a medium to realize the interaction between virtual reality and reality.In augmented reality system, tracking and registration technology is the basic technology to construct augmented reality system, and it is also the key to determine the performance of augmented reality system.The traditional tracking registration algorithm based on computer vision is complex and time-consuming.Compared with the computer, the current smart phone has less memory and less processing power, so it is not enough to run the algorithm with higher complexity.The accuracy and real-time of the augmented reality system on the mobile phone can not be compared with the augmented reality system on the PC side.Therefore, in view of the performance and functional requirements of mobile augmented reality technology, this paper compares the advantages and disadvantages of various feature point matching algorithms by comparing various feature point matching algorithms.ORB algorithm is selected as the feature point matching algorithm of the mobile augmented reality system.Based on the analysis of ORB algorithm, aiming at the problem that the feature points detected by ORB algorithm are sensitive to the scale change, an improved ORB algorithm is proposed, which combines the idea of scale space.In the phase of feature point matching and pose tracking, the traditional ORB algorithm uses exhaustive method for feature point matching, which requires a large amount of computation and is not suitable for the mobile phone.Therefore, the method of visual word bag model is introduced to change the feature search strategy, so as to accelerate the matching efficiency of feature points, calculate camera pose, and realize the real-time tracking of mobile phone pose.The mobile augmented reality system is designed and implemented, and the framework of mobile augmented reality system is proposed. The improved algorithm is applied to the system.In order to improve the robustness and real-time performance of mobile phone algorithm, mobile augmented reality system is suitable for large scale indoor and outdoor scenes.
【学位授予单位】：沈阳工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.9

【相似文献】