面向驾驶模拟的高速公路三维实景重现建模研究毕业论文
2021-06-08 00:19:48
摘 要
随着科技的发展,道路交通仿真已经逐步出现在人们的面前,汽车驾驶模拟器也成为现下一大热门。驾驶员对模拟视景的逼真程度的要求也越来越高,驾驶员希望能够从驾驶模拟器的操作中体验到真实的驾驶感受。通过对驾驶环境的实景建模,能够为驾驶员提供较为逼真的驾驶环境,使得驾驶员能够在虚拟驾驶中看到真实驾驶中的交通场景,景物和地形,能够使得驾驶员感觉身在真实的驾驶环境中一样。同时,在道路交通仿真中引入交通流模型和其他交通相关的技术,能够对高速公路的设计合理性进行分析与安全评价研究,在对驾驶员进行驾驶训练的时候,驾驶员如果在真实的驾驶环境中可能会产生许多的驾驶安全问题,但在驾驶模拟环境中却不会出现任何的安全隐患,驾驶员能够在逼真的驾驶环境中获得必须的驾驶技术。同时也能够比较方便的对驾驶员在驾驶过程中的行为和心理进行分析。
道路提取是视景建模中非常重要的一部分,它是从卫星遥感图像中提取道路特征的一种重要手段。道路提取是道路视景建模的一切的基础,没有道路提取就无法获得正确的道路信息,也就无法进行之后的道路建模工作。道路信息作为驾驶模拟体验环境的重要组成部分,能够对它进行识别和精确定位对逼真环境的构造有着及其重要的意义。然而高分辨率的道路提取的图像的信息量巨大,且存在许多干扰信息影响提取的精度,因此对道路进行高分辨率的道路提取是十分困难的,道路当中众多的多余信息和成像时的噪声会严重的影像道路提取的精确度。特别是对影像的噪声过滤不够足够或者是对影像的成像杂质没有良好的处理,最后可能会使得提取的结果产生极大的误差。
本文从数学形态学和边缘检测两种方法出发,对高速公路道路的遥感影像提取进行了一定的研究,总结了两种不同的道路提取方法,并且都能够很好的获得道路的平行边缘。
数学形态学是一门建立在严格的理论基础上的数字图像处理方法。它使得处理工具和待处理图像都被看作是集合,运用成熟的集合理论来对图像处理进行研究,使得形态学有着数学那极其严谨的特点。边缘检测对于道路有着明显的集合特征的时候能够有着比较好的研究,选择边缘检测算子的标准取决于算子需要应用的影像的自身,许多时候其实是不需要太复杂的高级算子就可以获得比较好的提取效果。
本文提取的效果在图像分辨率不高,道路的直线性比较好的情况下有着比较好的提取效果。而高速公路的直线性一般都比较好,因此本文的方法在高速公路的道路提取方面比较好
关键词:半自动 道路提取 边缘检测 数学形态学 图像增强
Abstract
When the vehicle driving simulator is more and more widely used in road traffic simulation, the drivers have a higher requirement for experiencing a vivid virtual environment. The drivers want to be able to experience the real driving experience from the operation of the driving simulator. By reproducing the modeling research on the highway of the three-dimensional , the driver can experience the more realistic driving environment, so the driver can in the virtual driving environment to see real traffic signs, scenery and terrain, and the driver feel the body is in the real driving environment. What’s more, after the traffic flow model and other traffic related technologies are introduced into the road traffic simulation , To analyze and evaluate the rationality of the design of Expressway is more easily. At the same time, the driver's psychological and behavioral characteristics can be studied in the safe condition.
Road extraction is a very important part of visual modeling. It is a process of extracting road features from remote sensing images of satellite. Road extraction is the basis of all the road scene modeling, there is no way to get the correct road information without the road extraction. As an important part of driving simulation experience environment, road information can be identified and precisely positioned to the realistic environment. However, High resolution road extraction of the image have a huge amount of information, and there are many interference information affecting the accuracy of the extraction, So it is very necessary to extract the road with high precision.
Mathematical morphology is a kind of digital image processing method based on strict theory. It makes the processing tools and the image to be treated as a collection, the use of sophisticated theory of the collection of image processing research, making the mathematical morphology has a very strict characteristics. Edge detection for the road has a clear set of features is able to have a relatively good study. The criteria for selecting the edge detection operator depends on the image of the operator to be applied. Many times, in fact, it is not need to be too complex advanced operator to get better extraction results.
In this paper, the effect of extraction in image resolution is not high, the road of a straight line is better to have a better extraction effect. And the straight line of the highway is generally better. Therefore, the method of this paper is relatively good in the road extraction of the highway.
Key words:Mathematical morphology;Road extraction;edge detection;
semi-automatic;image enhanceme
目录
摘要 I
Abstract III
目录 V
第1章绪论 1
1.1研究背景和意义 1
1.2遥感影像的研究现状及发展趋势 2
1.2.1遥感影像的道路特征 2
1.2.2半自动道路提取研究现状 3
1.2.3自动道路提取研究现状 5
1.2.4遥感影像道路提取发展趋势 7
1.3论文的主要内容和结构安排 7
第2章 基于数学形态学的道路提取 9
2.1结构元素 10
2.2二值形态学基本运算 10
2.2.1二值形态学的腐蚀和膨胀运算 10
2.2.2二值形态学的开,闭运算 12
2.2.3击中与击不中变换 13
2.3灰度形态学基本运算 13
2.3.1灰度腐蚀 13
2.3.2灰度膨胀运算 14
2.3.3灰度开运算和闭运算 14
2.4数学形态学在道路提取中的应用 14
2.4.1数学形态学的道路提取可行性分析 14
2.4.2数学形态学道路提取路线 15
2.4.3数学形态学道路提取 16
2.5提取结果评价 21
2.6本章小结 22
第3章 基于边缘检测的道路提取 23
3.1常见的边缘检测算子介绍 24
3.2利用边缘检测道路提取的研究路线 28
3.3图像预处理 28
3.3.1图像的增强处理 29
3.3.2图像的平滑处理 31
3.4边缘检测结果比较 34
3.5提取目标优化 36
3.6提取结果评价 38
3.7本章小结 38
第4章 总结和展望 39
参考文献 41
附录 44
致谢 48