顾及几何特征的数学形态学高分辨率遥感道路提取方法研究

发布时间：2018-02-03 07:18

本文关键词： 高分辨率遥感影像道路几何特征数学形态学道路提取　出处：《中南大学》2013年硕士论文　论文类型：学位论文

【摘要】：随着遥感技术和计算机技术的快速发展,高分辨率遥感为人们提供了高质量和丰富的数据源。道路是基础地理信息中的重要基础要素,道路网的高效率、高精度提取对地图制图、数据更新具有十分重要的意义,而高分辨率遥感技术已成为道路网信息获取的重要手段。道路信息在遥感影像中呈现出以下特征：宽度基本一致长度长的几何特征、灰度在局部区域变化缓慢的辐射特征、鲁棒性强的光谱特征、道路之间相互连接成网络的拓扑特征、道路在现实中所实现的功能特征以及道路与周围区域相关联的上下文特征。如何从丰富的高分辨率遥感数据中提取道路网是遥感目标提取的重要内容。数学形态学作为应用数学的一个重要分支,它是通过探测目标图像的形态特征来研究目标图像的空间结构从而对图像进行各种处理的分析工具,它的基本思想是用具有一定形态特征的结构元素去度量和提取图像中对应的形状目标以达到对图像分析和识别的目的。数学形态学在计算机视觉、图像处理和分析、模式识别、测绘等领域得到了广泛的应用。本文基于数学形态学的基础理论和算法,结合遥感影像中道路的几何特征对高分辨率遥感影像的道路网信息进行提取。本文的主要内容有： 1.简要叙述了道路网提取的意义,介绍了国内外数学形态学应用于图像处理和道路信息获取的研究概况。 2.介绍了数学形态学的理论和基本运算,主要对二值形态学和灰度形态学的理论、算法和代数性质做了阐述。 3.分析了高分辨率遥感影像道路提取的特殊性与难点,运用数学形态学开、闭运算,Top-Hat变换,阈值分割,形态细化等算法,在模拟影像上实现了道路信息的精确提取。 4.在对遥感影像中道路特征进行深入研究与分析的基础上,充分顾及道路的几何特征,把在模拟影像上准确实现道路提取的数学形态学算法应用于实验影像中的道路网提取。采用形态开运算对实验影像去噪；闭运算用于填补道路提取过程中的漏洞；利用阈值分割方法对影像进行分割；最后通过细化运算得到道路的中心线信息,进而得到完整的道路网提取结果。本项研究表明,在充分顾及遥感影像中道路几何特征的基础上,运用数学形态学方法在高分辨率遥感影像中提取道路,具有较好的提取效果和较高的几何精度。
[Abstract]:With the rapid development of remote sensing technology and computer technology, high-resolution remote sensing provides people with high quality and rich data sources. High precision extraction is very important to map mapping and data updating. High-resolution remote sensing technology has become an important means of road network information acquisition. Road information in remote sensing images show the following characteristics: the width of the geometric features of the basic length of the same length. The gray level changes slowly in the local region radiation characteristic, the robust spectrum characteristic, the road interconnects the network topology characteristic. How to extract road network from rich high-resolution remote sensing data is an important content of remote sensing target extraction. As an important branch of applied mathematics, mathematical morphology is an analytical tool to study the spatial structure of the target image by detecting the morphological characteristics of the target image. Its basic idea is to measure and extract the corresponding shape objects in the image by using the structural elements with certain morphological characteristics to achieve the purpose of image analysis and recognition. Mathematical morphology in computer vision. Image processing and analysis, pattern recognition, mapping and other fields have been widely used. This paper based on the basic theory and algorithm of mathematical morphology. The road network information of high resolution remote sensing image is extracted by combining the geometric characteristics of road in remote sensing image. The main contents of this paper are as follows: 1. The significance of road network extraction is briefly described, and the research situation of the application of mathematical morphology in image processing and road information acquisition at home and abroad is introduced. 2. The theory and basic operation of mathematical morphology are introduced. The theories, algorithms and algebraic properties of binary morphology and grayscale morphology are discussed. 3. The particularity and difficulty of road extraction in high resolution remote sensing image are analyzed, and the algorithms of mathematical morphology opening, closed operation and Top-Hat transform, threshold segmentation and morphological thinning are used. The accurate extraction of road information is realized on the simulated image. 4. Based on the in-depth study and analysis of road features in remote sensing images, the geometric features of roads are fully taken into account. The mathematical morphology algorithm is applied to the road network extraction in the experimental image, and the morphological open operation is used to de-noise the experimental image. Closed operation is used to fill the loophole in the road extraction process; The threshold segmentation method is used to segment the image. Finally, the central line information of the road is obtained by thinning operation, and the complete road network extraction result is obtained. This study shows that, on the basis of taking full account of the geometric characteristics of the road in remote sensing images, the method of mathematical morphology is used to extract roads in high-resolution remote sensing images. It has better extraction effect and higher geometric precision.
【学位授予单位】：中南大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：P237

【参考文献】