基于影像的室内定位方法研究

发布时间：2018-06-20 15:50

本文选题：室内定位 + 影像匹配　；参考：《西安科技大学》2017年硕士论文

【摘要】：在大数据时代,随着互联网技术和智能手机的快速发展,基于位置的移动服务越来越深入的进入各个行业和人们的日常生活中。据统计,人们80%的时间处在室内环境中,高精度的室内定位能够给人们带来工作和生活的便利,尤其是在博物馆、电影院以及商场等室内大型场馆。针对传统室内定位方法定位精度低、设备昂贵、便携性差等问题。本文提出了一种C/S(Client-Server)架构下基于影像的空间定位方法。该方法具有精度高、可视化效果好、信息丰富、便携性强等优点,成为近年来室内定位领域的研究热点之一。本文方法的实现主要分为客户端和服务端两部分,客户端主要负责调用手机内置相机获取室内定位影像,并将影像传输到服务端,在服务端中预先建立物方特征库,接收上传的定位影像,计算定位影像的外方位元素,然后将计算结果传回客户端,三维显示用户的位置,完成用户室内定位。本文主要研究内容和创新点如下:(1)、详细的介绍了针孔相机模型和基于共线条件方程描述的相机模型,分析了现有相机检校方法的数学模型和检校精度的优劣。利用西安科技大学室内三维相机检校控制场,获得相机内方位元素和镜头畸变参数。(2)、采用计算机视觉中SFM(structure from motion)方法计算像点对应的物方点坐标。并将序列影像中特征点及其像点坐标、物方点三维坐标整理成物方特征库,为求解定位影像外方位元素提供物方点坐标。(3)、定位影像外方位元素的计算。提取定位影像特征点并与物方特征库匹配,获取物方点坐标。通过最大似然估计法求出投影矩阵H,然后分解投影矩阵获得影像外方位元素初始值。最后通过共线方程精确求解影像的外方位元素。实验表明,本文中方法能够达到厘米级定位精度,满足用户定位精度的要求。(4)、将三维场景模型和相关代码打包成手机APP软件,实现调用手机摄像头拍摄定位影像和三维漫游显示定位结果的功能。
[Abstract]:In the era of big data, with the rapid development of Internet technology and smart phone, location-based mobile services are more and more deeply into every industry and people's daily life. According to statistics, people spend 80% of the time in the indoor environment, high precision indoor positioning can bring people work and life convenience, especially in museums, cinemas and shopping malls and other indoor large-scale venues. Aiming at the problems of low precision, expensive equipment and poor portability of traditional indoor positioning methods. In this paper, an image based spatial location method based on C / S Client-Server is proposed. This method has the advantages of high precision, good visualization effect, abundant information and strong portability, and has become one of the research hotspots in the field of indoor positioning in recent years. The realization of this method is mainly divided into two parts: client and server. The client is mainly responsible for calling the built-in camera of mobile phone to obtain the indoor positioning image, and transmitting the image to the server, and establishing the object side signature library in the server in advance. After receiving the uploaded location image and calculating the location elements of the location image, the result is sent back to the client, and the location of the user is displayed in three dimensions to complete the location of the user's room. The main contents and innovations of this paper are as follows: 1. This paper introduces the pinhole camera model and camera model based on collinear conditional equation in detail, and analyzes the mathematical model of the existing camera calibration methods and the advantages and disadvantages of calibration accuracy. By using the control field of 3D camera in Xi'an University of Science and Technology, the azimuth elements and lens distortion parameters of the camera are obtained, and the coordinates of the object points corresponding to the image points are calculated by using the SFM structure from motionof computer vision. The feature points and their coordinates in the sequence images and the three-dimensional coordinates of the object points are arranged into a feature library, which provides the coordinates of the object points and the coordinates of the location images for solving the external orientation elements of the location images, and the calculation of the external orientation elements of the location images. The feature points of the location image are extracted and matched with the object-side feature database to obtain the coordinates of the object-square points. The projection matrix H is obtained by the maximum likelihood estimation method, and then the initial value of the external azimuth element of the image is obtained by decomposing the projection matrix. Finally, the external azimuth element of the image is solved accurately by collinear equation. Experiments show that the method in this paper can achieve centimeter level positioning accuracy and meet the requirements of user positioning accuracy. The 3D scene model and related codes are packaged into mobile phone app software. The function of calling mobile phone camera to shoot location image and three-dimensional roaming display location result is realized.
【学位授予单位】：西安科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;P23

【参考文献】