非监督的SAR图像海陆分割与溢油提取方法研究
发布时间：2018-10-29 09:21
【摘要】：摘要： 海洋溢油污染不仅影响人类经济的发展,更影响到海洋生态环境,对海洋动植物、沿海人民的生活带来了巨大的威胁。及时、准确的监测海面溢油,对海洋生态环境保护以及人类经济的发展具有十分重要的意义。本文针对这一问题,立足于建立实时、无人监督的海洋监测系统,选用了成本低、全天时、全天候的SAR图像来进行海面溢油信息的监测,通过对无人监督海陆分割和暗区域(溢油)两方面的研究,找到了快速准确的方法,通过相应的实验验证了其可行性,同时开发完成了一体化SAR图像溢油检测框架系统。 论文主要论述了以下内容： 1.简单介绍了SAR的基本原理及图像特点,介绍SAR图像暗区域(溢油)的研究背景及意义,并分析了当前国内外的常用方法及存在的不足,同时介绍了本文的研究思路。 2.首先介绍常见的海陆方法及原理,同时针对实现无人监督的海陆分割,基于C-V模型的水平集方法进行了探讨,并给出了改进方案实现了针对于高分辨率SAR数据的应用,同时将实验结果与原模型进行比较,证明了该方案的可行性。 3.从溢油成像原理和成像条件进行阐述,介绍了SAR图像油膜的特征,并进一步对海面暗区域(溢油)自动分割方法进行探讨,利用开窗分类将单纯的分割问题转化为分类问题,同时通过快速的CFAR算法提取每一类中的暗区域,通过详细的实验证明了方法的可行性。最后介绍了基于SAR图像的一体化溢油检测框架。 4.总结了当前溢油检测方法的优缺点,以及未来研究的方向。
[Abstract]:Abstract: Marine oil spill pollution not only affects the development of human economy, but also affects the marine ecological environment, and poses a great threat to marine animals and plants and the lives of coastal people. Timely and accurate monitoring of oil spill on the sea level is of great significance to the protection of marine ecological environment and the development of human economy. Aiming at this problem, based on the establishment of a real-time, unsupervised ocean monitoring system, this paper selects the low cost, all-weather, all-weather SAR images to monitor the oil spill information on the sea surface. Through the research of unsupervised sea and land segmentation and dark area (oil spill), a fast and accurate method is found, and the feasibility is verified by corresponding experiments. At the same time, an integrated SAR image oil spill detection framework system is developed. This paper mainly discusses the following contents: 1. This paper briefly introduces the basic principle and image characteristics of SAR, introduces the research background and significance of dark area (oil spill) in SAR image, analyzes the common methods and shortcomings at home and abroad, and introduces the research ideas of this paper. 2. Firstly, the common methods and principles of land and sea are introduced. At the same time, the level set method based on C-V model is discussed in order to realize the unsupervised land and sea segmentation, and the improved scheme is given to realize the application of high-resolution SAR data. At the same time, the experimental results are compared with the original model, and the feasibility of the scheme is proved. 3. This paper expounds the principle and condition of oil spill imaging, introduces the characteristics of oil film in SAR image, and further discusses the automatic segmentation method of sea surface dark area (oil spill), and transforms the simple segmentation problem into classification problem by using windowed classification. At the same time, a fast CFAR algorithm is used to extract the dark areas in each class, and the feasibility of the method is proved by detailed experiments. Finally, an integrated oil spill detection framework based on SAR image is introduced. 4. The advantages and disadvantages of current oil spill detection methods and the future research directions are summarized.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN957.52
本文编号：2297286
[Abstract]:Abstract: Marine oil spill pollution not only affects the development of human economy, but also affects the marine ecological environment, and poses a great threat to marine animals and plants and the lives of coastal people. Timely and accurate monitoring of oil spill on the sea level is of great significance to the protection of marine ecological environment and the development of human economy. Aiming at this problem, based on the establishment of a real-time, unsupervised ocean monitoring system, this paper selects the low cost, all-weather, all-weather SAR images to monitor the oil spill information on the sea surface. Through the research of unsupervised sea and land segmentation and dark area (oil spill), a fast and accurate method is found, and the feasibility is verified by corresponding experiments. At the same time, an integrated SAR image oil spill detection framework system is developed. This paper mainly discusses the following contents: 1. This paper briefly introduces the basic principle and image characteristics of SAR, introduces the research background and significance of dark area (oil spill) in SAR image, analyzes the common methods and shortcomings at home and abroad, and introduces the research ideas of this paper. 2. Firstly, the common methods and principles of land and sea are introduced. At the same time, the level set method based on C-V model is discussed in order to realize the unsupervised land and sea segmentation, and the improved scheme is given to realize the application of high-resolution SAR data. At the same time, the experimental results are compared with the original model, and the feasibility of the scheme is proved. 3. This paper expounds the principle and condition of oil spill imaging, introduces the characteristics of oil film in SAR image, and further discusses the automatic segmentation method of sea surface dark area (oil spill), and transforms the simple segmentation problem into classification problem by using windowed classification. At the same time, a fast CFAR algorithm is used to extract the dark areas in each class, and the feasibility of the method is proved by detailed experiments. Finally, an integrated oil spill detection framework based on SAR image is introduced. 4. The advantages and disadvantages of current oil spill detection methods and the future research directions are summarized.
【学位授予单位】：北京交通大学
【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10条
1 黄晓霞,朱振海;海洋表面油膜的合成孔径雷达图像检测[J];地质科技情报;1997年S1期
2 姜慧研;冯锐杰;;基于改进的变分水平集和区域生长的图像分割方法的研究[J];电子学报;2012年08期
3 徐青;郑汲;程永存;纪棋严;;应用纹理分析识别SAR海上溢油图像[J];河海大学学报(自然科学版);2011年05期
4 徐玲凌;肖进胜;易本顺;娄利军;;改进的C-V水平集模型图像分割算法[J];计算机工程;2009年10期
5 官晟,吴东,张博,刘智深;海洋激光荧光雷达海面油膜种类鉴别研究[J];青岛海洋大学学报(自然科学版);2002年02期
6 石立坚;赵朝方;刘朋;;基于纹理分析和人工神经网络的SAR图像中海面溢油识别方法[J];中国海洋大学学报(自然科学版);2009年06期
7 李建明;海洋石油污染的危害与净化[J];生物学教学;2002年07期
8 刘书君;冉强军;袁运能;毛士艺;;基于单通道合成孔径雷达子图像的动目标检测性能分析[J];系统工程与电子技术;2010年12期
9 赵冬至,丛丕福;海面溢油的可见光波段地物光谱特征研究[J];遥感技术与应用;2000年03期
10 原君娜;邵芸;田维;王世昂;;利用SAR图像识别海面油膜的方法介绍[J];遥感技术与应用;2010年01期
，本文编号：2297286
本文链接：https://www.wllwen.com/kejilunwen/wltx/2297286.html
