SAR图像配准以及变化检测的研究

【作者】 丁亚兰；

【导师】 王建国；

【作者基本信息】 电子科技大学 ， 信号与信息处理， 2010， 硕士

【摘要】 合成孔径雷达(SAR)具有全天候、全天时的特点,可方便地获得同一地区不同时段的图像。SAR图像变化检测技术通过对不同时段SAR图像的综合分析,检测同一场景是否有变化发生。SAR图像变化检测已经应用于很多方面,如对人工检测比较困难的热带雨林、沙漠等自然条件恶劣的地区进行实时监测,以了解生态环境变化的情况;对农田进行监测,分析农作物生长状况;对城区环境进行监控,合理规划城市布局,规范土地的使用;对军事目标进行监测,了解兵力部署、军事调动等战场势态,并可用于打击效果评估。本文研究了合成孔径雷达图像变化检测及相关技术。首先,介绍SAR图像基本特性。总结了SAR图像与光学图像的区别、SAR图像的统计分布特性;分析了不同SAR图像之间的几何特性,包括平移、缩放、旋转等;阐述了相干斑产生的机理,通过试验比较了几种经典的抑制SAR图像相干斑的算法。其次,分析SAR图像配准技术。给出图像配准的定义、一般步骤;分析了图像的仿射变换模型;系统归纳了图像配准的常见算法和配准精度评估方法;结合图像能量分布改进提出了一种基于能量分布的SAR图像配准算法;对于强散射目标不明显的SAR图像,本文提出了一种基于不变矩的自动配准法。实验结果证实,这两种方法具有较好的算法稳健性,配准精度较高。最后,研究SAR图像变化检测算法。介绍SAR图像变化检测的基本过程;总结图像变化检测的基本方法和评估检测效果的一般方法;改进提出了一种基于子图像的主分量分析变化检测算法;另外,对于主次分量不明显的SAR图像,本文提出了一种基于最大似然估计变化检测算法。这两种方法适用范围更广,能够取得很好的检测效果。更多还原

【Abstract】 Change detection is a process that analyzes a pair of images acquired on the same geographical area at different times in order to identify changes that may have occurred between the considered acquisition dates. The major advantage of synthetic aperture radar (SAR) is its all-weather capability that allows the acquisition of time series of imagery with exact acquisition dates under all climatic conditions. This paper research change detection technology and the related technology using synthetic aperture radar image.Firstly, the basic characteristics of SAR image are introduced, including the differences between SAR images and optical images, the geometric characteristics,statistical distribution,speckle noise of SAR images, comparing several classic de-speckle noise algorithms by experiments.Secondly, The definition of image registration is given, the general steps; combination of improved energy distribution of the image is presented based on energy distribution of the SAR image registration; obvious target for the strong scattering of SAR images is proposed based on moment invariant of the automatic registration method. The results confirmed that both methods have good stability of the algorithm, registration accuracy.Finally, We study the SAR image change detection algorithm. Summarize the basic method of image change detection and assessment of the general method of detecting effect; improvement is presented based on principal component analysis for sub-image change detection algorithm; In addition, this paper proposes a maximum likelihood estimation based on change detection algorithm. Wider application of these two methods can achieve good test results.更多还原