论文总字数:28296字
摘 要
论文作者签名: 导师签名: 日期: 2018年6月1日
基于高分辨率遥感影像的道路信息提取方法
摘 要
现今,日新月异的城市景观对实时获取城市地理信息提出了更高的要求。道路网作为城市建设中的基础地理信息,在交通规划、土地利用等领域具有重要的意义。传统中低分辨率的遥感影像提供的道路信息有限,而高分辨率遥感卫星的发展为道路的精确提取提供了可能。高分辨率遥感影像在带来详尽地理信息的同时,也造成了道路特征提取的困难——图像细节引起的噪声使提取出的道路不连续。
本文主要研究的是基于高分辨率遥感影像的城市区域的道路提取方法。首先,通过阅读大量文献和书籍,本文总结整理了国内外研究现状和遥感图像处理的基础理论。其次,本文选用了两幅空间分辨率小于5m的城市区域高分辨率遥感影像,对两种现有的道路提取方法进行测试。最后,本文对测试结果进行了评估,并对未来的研究方向作出了设想。
在Hough变换的实验中,提取目标是道路边缘。本文通过MATLAB完成了传统Hough变换与改进Hough变换的算法测试,表明二者在提取精度差异不大的情况下,改进后的Hough变换完成了算法效率上的优化。在面向对象的实验中,提取目标是道路中心线。本文运用了ERDAS的内置模块ERDAS IMAGINE Objective进行算法测试,该模块丰富的算子为建立道路提取模型提供了良好的技术支撑,但多个阈值选取的复杂性使该方法的适用性不强。最终结果表明,高分辨率遥感影像的复杂性使传统的道路提取方法具有一定局限性,因此,该方向的研究应偏向于将多种方法应用在道路提取的各个阶段,克服传统方法的局限性。
关键词:高分辨率遥感影像,道路提取,Hough变换,图像处理,面向对象的方法,MATLAB,ERDAS IMAGINE Objective
Road information extraction method based on
high-resolution remote sensing images
Abstract
Nowadays,the ever-changing urban landscape has higher requirements for real- time access to urban geographic information.As the basic geographical information in urban construction, road network is of great significance in the fields of traffic planning and land use.The traditional medium-resolution and low-resolution remote sensing images provide limited road information,and the development of high-resolution remote sensing satellites provides the possibility for accurate road extraction.High resolution remote sensing image not only brings detailed geographical information,but also makes it difficult to extract road features.
This paper mainly studies the road extraction method of urban areas based on high-resolution remote sensing images.First of all,through reading a large number of literature and books,this paper summarizes the domestic and foreign research status and basic theories of remote sensing image processing.Secondly,two high-resolution remote sensing images of urban areas with spatial resolution less than 5m were selected to test two existing road extraction methods.Finally,the test results are evaluated and the future research directions are envisaged.
In the experiment of Hough transformation,the extraction target is the road edge. By testing the traditional Hough transform and improved Hough transform algorithm based on the MATLAB,it shows the improved Hough transform complete the efficiency optimization algorithm when two methods show little difference in extraction accuracy.In object-oriented experiments,the extraction target is the road central line.This paper uses the ERDAS built-in module ERDAS IMAGINE Objective to test the algorithm,the module’s abundant operator for road extraction model provides a good technical support,but the complexity of multiple threshold selecting makes the applicability of this method be not strong.The final results show that the complexity of high resolution remote sensing image makes traditional road extraction methods have a certain limitation,as a result,the direction of research should be geared towards applying variety of methods to each stages of road extraction,to overcome the limitations of traditional methods.
KEY WORDS: High-resolution remote sensing images,road extraction,Hough transform,image processing,object-oriented methods,MATLAB,ERDAS IMAGINE Objective.
目 录
摘 要 I
Abstract II
第一章 绪论 1
1.1研究背景与意义 1
1.2国内外研究现状 1
1.2.1基于图像分割的道路提取方法 1
1.2.2模板匹配法 2
1.2.3 基于边缘的道路提取方法 2
1.2.4特定模型法 2
1.2.5多分辨率分析法 3
1.2.6面向对象的方法 3
1.3研究内容与技术路线 4
1.4论文结构 5
第二章 基于高分辨率遥感影像的道路信息提取基础理论与技术 6
2.1道路分级及特征 6
2.1.1道路分级 6
2.1.2道路的特征 6
2.2道路提取的技术基础 8
2.2.1图像特征理论 8
2.2.2图像增强 8
2.2.3图像分割 10
第三章 基于高分辨率遥感影像的道路信息提取实验过程与分析 13
3.1实验装置 13
3.1.1数据描述 13
3.1.2精度评估标准 13
3.2实验过程 13
3.2.1Hough变换提取道路 13
3.2.2面向对象的提取方法 19
第四章 研究总结与展望 26
致谢 27
参考文献 28
第一章 绪论
1.1研究背景与意义
剩余内容已隐藏,请支付后下载全文,论文总字数:28296字
该课题毕业论文、开题报告、外文翻译、程序设计、图纸设计等资料可联系客服协助查找;