基于多目标聚类与非局部均值的SAR图像变化检测
发布时间：2019-06-22 14:44
【摘要】：最近这些年来,随着遥感技术的发展,合成孔径雷达因其成像不受光照的影响,越来越受到人们的关注,特别是在变化检测领域。变化检测是指让计算机检测出同一个地方不同时间的两幅图像之间的差异。本文对合成孔径雷达图像变化检测问题进行了研究,提出了基于多目标聚类与非局部均值的合成孔径雷达图像变化检测的方法,所取得的主要研究成果有以下两个方面:1.提出了一种基于多目标聚类的差异图分析方法。众所周知,图像中同时存在图像细节和图像噪声,所以在变化检测任务中权衡细节的保留和噪声的不敏感扮演了一个关键的角色。本文定义了两个目标函数,一个目标函数表示保持图像细节,另一个目标函数表示滤除图像噪声。这样就把变化检测问题转变为多目标优化的问题,然后再采用多目标进化算法对目标函数进行优化。作为结果,我们可以获得一组有着不同程度的最优解,用户可以选择针对特定问题的合适的解。2.提出了一种基于非局部均值的多目标聚类的变化检测方法。在图像去噪中,相较于局部滤波器,非局部滤波器表现出更好的性能。本文使用非局部均值对差异图去噪表示滤出噪声的目标函数中的图像。非局部均值中块与块之间的相似度度量是一个关键的问题。为了更适应合成孔径雷达图像,比值相似度度量被应用于非局部均值算法中。
[Abstract]:In recent years, with the development of remote sensing technology, synthetic aperture radar (SAR) has attracted more and more attention because its imaging is not affected by light, especially in the field of change detection. Change detection refers to the computer to detect the difference between two images at different times in the same place. In this paper, the problem of SAR image change detection is studied, and a method of SAR image change detection based on multi-target clustering and non-local mean is proposed. The main research results are as follows: 1. A difference graph analysis method based on multi-objective clustering is proposed. As we all know, there are both image details and image noise in the image, so it plays a key role in weighing the retention of details and the insensitivity of noise in the task of change detection. In this paper, two objective functions are defined, one to maintain image details and the other to filter out image noise. In this way, the change detection problem is transformed into the multi-objective optimization problem, and then the multi-objective evolutionary algorithm is used to optimize the objective function. As a result, we can obtain a set of optimal solutions with different degrees, and the user can choose the appropriate solution for a particular problem. 2. A change detection method based on non-local mean value for multi-objective clustering is proposed. Compared with the local filter, the nonlocal filter shows better performance in image denoising. In this paper, the nonlocal mean is used to Denoise the difference graph to represent the image in the objective function of filtering noise. The measure of similarity between blocks in nonlocal mean is a key problem. In order to be more suitable for synthetic aperture radar (SAR) images, the ratio similarity measure is applied to the nonlocal mean algorithm.
【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
[Abstract]:In recent years, with the development of remote sensing technology, synthetic aperture radar (SAR) has attracted more and more attention because its imaging is not affected by light, especially in the field of change detection. Change detection refers to the computer to detect the difference between two images at different times in the same place. In this paper, the problem of SAR image change detection is studied, and a method of SAR image change detection based on multi-target clustering and non-local mean is proposed. The main research results are as follows: 1. A difference graph analysis method based on multi-objective clustering is proposed. As we all know, there are both image details and image noise in the image, so it plays a key role in weighing the retention of details and the insensitivity of noise in the task of change detection. In this paper, two objective functions are defined, one to maintain image details and the other to filter out image noise. In this way, the change detection problem is transformed into the multi-objective optimization problem, and then the multi-objective evolutionary algorithm is used to optimize the objective function. As a result, we can obtain a set of optimal solutions with different degrees, and the user can choose the appropriate solution for a particular problem. 2. A change detection method based on non-local mean value for multi-objective clustering is proposed. Compared with the local filter, the nonlocal filter shows better performance in image denoising. In this paper, the nonlocal mean is used to Denoise the difference graph to represent the image in the objective function of filtering noise. The measure of similarity between blocks in nonlocal mean is a key problem. In order to be more suitable for synthetic aperture radar (SAR) images, the ratio similarity measure is applied to the nonlocal mean algorithm.
【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
【相似文献】
相关期刊论文 前10条
1 沈国键，夏启明，李瑛;基于知识自适应图像平滑[J];武汉测绘科技大学学报;1995年02期
2 林方特;;几种常见图像平滑技术的研究[J];才智;2009年04期
3 杨朝辉;张大鹏;李乃民;;人体生理图像与病理舌纹图像[J];哈尔滨工业大学学报;2009年12期
4 郭琦;;图像平滑的2维小波插值方法[J];中国图象图形学报;2010年10期
5 杨平先,孙兴波;基于粗集方法改进的图像平滑[J];电讯技术;2003年03期
6 王洪亮,曹苏明,刘国平,吴建华;改进的梯度倒数加权算法在图像平滑中的应用[J];红外技术;2003年04期
7 郭海涛,田坦,王连玉,闫宏生;利用粗集理论的声呐图像平滑[J];海洋技术;2005年02期
8 吕振肃,魏弘博,刘R，
本文编号：2504676
本文链接：https://www.wllwen.com/kejilunwen/wltx/2504676.html
