基于活动轮廓模型的SAR图像目标轮廓提取研究
发布时间：2019-02-18 14:45
【摘要】：合成孔径雷达(synthetic aperture radar, SAR)相较于传统的遥感探测方法具有全天时全天候的优点，广泛应用于军用和民用领域。目标轮廓表征了目标形状，可用于目标方位角估计、自动目标识别等，因此SAR图像目标轮廓提取任务受到了人们的广泛关注。受SAR图像相干斑噪声的影响，传统的目标轮廓提取方法应用在SAR图像时失效。为此，本文以活动轮廓模型为基础，充分考虑SAR图像中相干斑噪声的影响，研究SAR图像目标轮廓提取的方法。 本文首先对SAR图像数据特性进行了分析，在此基础上介绍了目前经典的SAR图像目标轮廓提取方法，，并通过实验证明了经典方法的缺点。 其次，结合SAR图像特点，对活动轮廓模型进行改造，基于向量场卷积模型与区域竞争模型进行结合的思路和基于分割后图像具有最大对比度的思路，提出了两种适用于SAR图像的目标轮廓提取方法。 再次，对活动轮廓模型初始轮廓的设置进行了研究，提出了适用于SAR图像目标轮廓提取的初始轮廓自动选取方法。 最后，基于仿真和实测SAR图像数据，对本文所研究的方法进行了验证。实验结果显示，本文所研究的两种目标轮廓提取方法能对SAR图像中目标轮廓进行准确、快速的提取，性能上明显优于现有的经典方法；本文研究的初始轮廓自动选取方法能显著减少活动轮廓模型的迭代次数，是一种有效的初始轮廓选取方法。
[Abstract]:The synthetic aperture radar (SAR) phase has the advantages of all-weather, and is widely used in military and civil fields. The target profile is characterized by the target shape, can be used for target azimuth estimation, automatic target recognition and the like, so that the target contour extraction task of the SAR image is widely concerned by the people. Due to the influence of the speckle noise of the SAR image, the traditional target contour extraction method is applied to the SAR image. This paper, based on the active contour model, takes full consideration of the influence of the speckle noise in the SAR image, and studies the method of the SAR image target profile extraction. In this paper, the characteristics of SAR image data are analyzed. On the basis of this, the method of extracting the target contour of SAR image is introduced, and the shortage of the classical method is proved by the experiment. Secondly, combining with the characteristics of SAR image, the active contour model is modified, and based on the combination of the vector field convolution model and the regional competition model and the idea of the maximum contrast based on the divided image, two target profiles suitable for the SAR image are proposed. In this paper, the initial contour of the active contour model is studied, and the initial contour from the target profile of the SAR image is proposed. The dynamic selection method is based on the simulation and the measured SAR image data. The results show that the two target contour extraction methods can be used for accurate and rapid extraction of the target profile in the SAR image, and the performance is obviously superior to that of the target contour. In this paper, the initial contour auto-selection method can significantly reduce the number of iterations of the active contour model, which is an effective method.
【学位授予单位】：南京航空航天大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN957.52
本文编号：2425946
[Abstract]:The synthetic aperture radar (SAR) phase has the advantages of all-weather, and is widely used in military and civil fields. The target profile is characterized by the target shape, can be used for target azimuth estimation, automatic target recognition and the like, so that the target contour extraction task of the SAR image is widely concerned by the people. Due to the influence of the speckle noise of the SAR image, the traditional target contour extraction method is applied to the SAR image. This paper, based on the active contour model, takes full consideration of the influence of the speckle noise in the SAR image, and studies the method of the SAR image target profile extraction. In this paper, the characteristics of SAR image data are analyzed. On the basis of this, the method of extracting the target contour of SAR image is introduced, and the shortage of the classical method is proved by the experiment. Secondly, combining with the characteristics of SAR image, the active contour model is modified, and based on the combination of the vector field convolution model and the regional competition model and the idea of the maximum contrast based on the divided image, two target profiles suitable for the SAR image are proposed. In this paper, the initial contour of the active contour model is studied, and the initial contour from the target profile of the SAR image is proposed. The dynamic selection method is based on the simulation and the measured SAR image data. The results show that the two target contour extraction methods can be used for accurate and rapid extraction of the target profile in the SAR image, and the performance is obviously superior to that of the target contour. In this paper, the initial contour auto-selection method can significantly reduce the number of iterations of the active contour model, which is an effective method.
【学位授予单位】：南京航空航天大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN957.52
【参考文献】
相关期刊论文 前7条
1 张俊梅;;SAR图像分割方法浅议[J];电脑知识与技术;2011年03期
2 周鑫;;基于EM算法的G0分布参数最大似然估计[J];电子学报;2013年01期
3 靳宏磊;朱蔚萍;李立源;陈维南;;二维灰度直方图的最佳分割方法[J];模式识别与人工智能;1999年03期
4 殷德奎,张保民,柏连发,王钰;一种热图像的多模板边缘检测方法[J];南京理工大学学报;1999年01期
5 成功;赵巍;毛士艺;;双阈值CFAR SAR目标分割方法[J];信号处理;2007年01期
6 金立左,夏良正,杨世周;图象分割的自适应模糊阈值法[J];中国图象图形学报;2000年05期
7 贺志国;陆军;匡纲要;;基于全局活动轮廓模型的SAR图像分割方法[J];自然科学进展;2009年03期
本文编号：2425946
本文链接：https://www.wllwen.com/kejilunwen/wltx/2425946.html
