立体影像景深修正方法及其在农村公路障碍物信息采集中的应用

发布时间：2018-01-03 22:13

本文关键词：立体影像景深修正方法及其在农村公路障碍物信息采集中的应用　出处：《昆明理工大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着我国社会经济水平和科技水平的不断提升,带来道路建设事业持续推进和交通运输信息化水平不断提高。如何更快、更廉价、更便捷的采集、更新道路信息,以用于道路的维护、保养以及智能决策,得到越来越多的关注。尤其是对于农村公路的维护保养,还需要面对数量庞大、维护人员少、维护保养经费低,同时又对广大人民的生活有重大影响的现实。那么如何构建适用于农村公路的数据采集系统,也就成了从事道路工程研究工作者必须关注的一个问题。论文以自动采集会影响农村公路正常通行状况的障碍物数据为切入点,探索利用立体影像景深修正方法采集农村公路道路信息的可能性。主要创新工作如下:1.提出了利用立体影像景深修正方法采集道路障碍物信息的解决方案。通过对现有道路信息采集系统的对比分析,结合农村公路的特点、前期研究,以及立体影像、深度图像的特点,提出通过深度图像修正双目图像获取障碍物空间信息,GPS进行定位的障碍物信息自动采集方案。2.提出了并实现了基于立体影像景深修正方法。采集并存储深度图像和双目图像,利用深度相机采集景深数据不易受背景光线影像的特性,深度图像和双目图像信息采集、处理,以及异源匹配和图像识别,实现了立体影像的深度修正。3.搭建了基于立体影像景深修正方法的农村公路障碍物信息自动采集原型系统,并进行了案例验证。通过组合GPS、深度相机及双面相机,基于EmguCV计算机视觉库,在Visual Studio 2015下采用C#编程语言开发集成了农村公路障碍物自动信息采集原型系统,并利用一个简单案例验证了其功能。系统的成功搭建以及案例验证表明,研究所提研究方案技术上可行,经济上合理。
[Abstract]:With the development of social economy and science and technology in our country, it brings about the continuous advancement of road construction and the continuous improvement of traffic and transportation informatization. How to collect more quickly, cheaply and conveniently. More and more attention has been paid to updating road information for road maintenance, maintenance and intelligent decision-making. Especially for the maintenance of rural roads, there are still a large number of maintenance staff. At the same time, it has a significant impact on the lives of the people. So how to build a data acquisition system suitable for rural roads. It has become a problem that the road engineering researchers must pay attention to. The paper takes the obstacle data which will affect the normal traffic condition of the rural highway as the breakthrough point. To explore the possibility of collecting rural highway road information by using stereo image depth of field correction method. The main innovative work is as follows:. 1. The solution of collecting road obstacle information by using stereo image depth correction method is put forward, and the comparison and analysis of the existing road information acquisition system are given. Combined with the characteristics of rural highway, the previous research, as well as the stereo image, the characteristics of depth image, it is proposed that the binocular image of depth image can be used to obtain the spatial information of obstacles. The automatic acquisition scheme of obstacle information based on GPS. 2. A method of depth correction based on stereo image is proposed and realized. The depth image and binocular image are collected and stored. Using depth camera to collect depth of field data is not easily subject to the characteristics of background light image, depth image and binocular image information acquisition, processing, as well as heterologous matching and image recognition. Based on the method of stereo image depth correction, a prototype system of automatic acquisition of obstacle information of rural highway is built, and the case is verified. Through the combination of GPS, this paper builds a prototype system of road obstacle information automatic acquisition based on the method of stereo image depth correction. Depth camera and dual camera, based on EmguCV computer vision library. Under Visual Studio 2015, the prototype system of automatic information collection for rural road obstacles is developed and integrated with C # programming language. A simple case is used to verify its function. The successful construction of the system and the case verification show that the proposed research scheme is technically feasible and economically reasonable.
【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U495

【参考文献】