基于马尔科夫随机场的SAR图像处理

发布时间：2018-03-01 03:05

本文关键词： SAR图像处理道路检测海陆分割马尔科夫随机场　出处：《西安电子科技大学》2014年硕士论文　论文类型：学位论文

【摘要】：SAR图像处理在各个领域已经得到广泛应用,是人们获取信息的重要途径之一。随着SAR成像技术的日益成熟,SAR数据的获取速度与质量也已得到迅速发展,这些都将增加对SAR图像解译的需求,在大量高分辨率SAR实测数据的支持下,实现快速、自动化的SAR图像处理是一个重要的研究方向。本文从概率角度出发,以马尔科夫随机场模型为基础,结合SAR图像数据的统计特性,实现图像中的目标检测与分割。论文主要工作总结如下:1.介绍了SAR图像处理中道路检测和海陆分割的研究背景、发展现状及意义。概括了论文的主要工作。2.介绍了马尔科夫随机场模型及基础理论。对于场的概念,邻域定义与子团的定义以及MRF(Markov Random Field)模型进行了详细的介绍。在应用马尔科夫随机场时最重要的理论支撑就是MRF与吉布斯随机场的等价性,本章对吉布斯随机场进行了介绍并对两者的等价性进行了介绍。最后介绍了常用的能量最小化算法。3.研究了SAR图像中海陆分割的算法的实现。采用合适的概率模型对SAR数据中海面和陆地两类样本的统计特性进行描述,得到两类样本的似然信息。然后结合马尔科夫随机场中的邻域与子团定义,将样本的先验信息考虑在内,通过能量函数的定义以及能量最小化算法完成海陆分割。4.研究了SAR图像中道路检测的算法实现。运用边缘检测完成图像中疑似道路的点检测,通过Hough变换将边缘检测的结果转换为线,实现由点到线的转换。SAR图像中的道路有着特定的属性,将这些属性作为道路的先验信息,借助马尔科夫随机场完成道路检测。实验结果表明了检测效果与场景的复杂度有着很大的关系,合理有效的预处理非常重要。
[Abstract]:SAR image processing has been widely used in various fields, and it is one of the important ways for people to obtain information. With the development of SAR imaging technology, the acquisition speed and quality of SAR data have been developed rapidly. All these will increase the demand for SAR image interpretation. With the support of a large number of high-resolution SAR measured data, it is an important research direction to realize fast and automatic SAR image processing. Based on Markov random field model and combined with the statistical characteristics of SAR image data, target detection and segmentation are realized. The main work of this paper is summarized as follows: 1. The research background of road detection and land and sea segmentation in SAR image processing is introduced. The main work of this paper is summarized. 2. Markov random field model and basic theory are introduced. The definition of neighborhood, the definition of sub-cluster and the MRF(Markov Random field model are introduced in detail. The most important theoretical support in applying Markov random field is the equivalence of MRF and Gibbs random field. In this chapter, Gibbs random field is introduced and its equivalence is introduced. Finally, the commonly used energy minimization algorithm .3.The realization of land and sea segmentation algorithm in SAR image is studied. To describe the statistical characteristics of sea and land samples in SAR data. The likelihood information of two kinds of samples is obtained, and the priori information of samples is taken into account by combining the definitions of neighborhood and sub-cluster in Markov random fields. Through the definition of energy function and the energy minimization algorithm to complete the land and sea segmentation. 4. The algorithm of road detection in SAR image is studied. The edge detection is used to detect the suspected road points in the image. The result of edge detection is transformed into a line by Hough transform. The path in the image is converted from point to line. The road has special attributes, which are regarded as the priori information of the road. The experimental results show that the detection effect is closely related to the complexity of the scene, and the reasonable and effective pretreatment is very important.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN957.52

【共引文献】