海面溢油数值模拟及其遥感图像处理的研究

发布时间：2018-02-23 06:34

本文关键词： 溢油海面 Marangoni溢油理论滤波图像分割　出处：《江苏科技大学》2014年硕士论文　论文类型：学位论文

【摘要】：随着社会的发展，人们生活水平的提高，石油资源不断被人们开发利用，海上船舶石油运输量急速增长，导致油井井喷和油轮溢油事故频繁发生，海洋水污染问题日趋严重。在海洋污染中，溢油事故发生频率高，不仅对人们的生产和生活造成严重危害，也对环境造成了一定的破坏。因此，尽快高效地解决溢油问题是非常必要的。溢油事故发生后，及时对海面进行数值模拟以及遥感图像处理都是很重要的。本论文在分析研究深水环境下的数值模拟的基础上，提出了一种适合于浅海环境下溢油海面的仿真方法以及提出一种新的适用于溢油图像的分割方法。本论文研究的主要内容如下： (1)首先从物理角度分析溢油的特性，在考虑水深因素的条件下，提出一种适合于浅海环境下溢油海面的仿真方法。首先利用TMA谱模型和Marangoni溢油理论模型，计算有限水深下溢油海面的海浪谱；然后根据海浪的色散关系和Longuet-Higgins海浪模型，计算有限水深下溢油海面的铅直位移；并分析了Kitaigorodskii深度函数的特性。 (2)本文其次对溢油图像进行滤波预处理，分析Lee滤波、Kuan滤波、Gamma-Map滤波、增强的Lee滤波算法以及增强的Kuan滤波算法对溢油图像的预处理效果，在此基础上，提出改进的增强Lee滤波方法，将增强的Lee滤波算法与伽马变换相结合，此方法能够在滤除相干斑噪声的同时并且能够保持图像的边缘信息及纹理特征。 (3)最后提出一种新的适应溢油图像的分割方法。图像分割方法是从图像中提取信息的一种手段，，是对图像进行分析的前提。本文通过对模糊C均值算法进行改进，利用图像的直方图作为此算法的初始值，克服了此算法对初始值敏感的缺点，能较准确地分割出溢油图像的溢油部分。
[Abstract]:With the development of society and the improvement of people's living standard, oil resources are being exploited and utilized constantly, and the oil transportation volume of offshore ships is increasing rapidly, which leads to oil well blowout and oil spill accidents occurring frequently. The marine water pollution problem is becoming more and more serious. In the marine pollution, oil spill accidents occur frequently, which not only cause serious harm to people's production and daily life, but also cause certain damage to the environment. It is very necessary to solve the problem of oil spill efficiently as soon as possible, and it is very important to simulate the sea surface and process remote sensing image in time after the oil spill accident. This thesis is based on the analysis and study of the numerical simulation in the deep water environment. A simulation method for oil spill surface in shallow sea environment and a new method for oil spill image segmentation are proposed. The main contents of this thesis are as follows:. 1) firstly, the characteristics of oil spill are analyzed from the physical point of view, and a simulation method suitable for oil spill surface in shallow water environment is proposed under the consideration of water depth factors. Firstly, the TMA spectrum model and Marangoni oil spill theory model are used. The wave spectrum of the oil spill surface with limited depth is calculated, and then the vertical displacement of the oil spill surface with limited depth is calculated according to the dispersion relation of the wave and the Longuet-Higgins wave model, and the characteristics of the Kitaigorodskii depth function are analyzed. Secondly, the Lee filter Kuan filter Gamma-Map filter, the enhanced Lee filter algorithm and the enhanced Kuan filter algorithm for oil spill image preprocessing are analyzed. An improved enhanced Lee filtering method is proposed, which combines the enhanced Lee filtering algorithm with the gamma transform. This method can not only filter speckle noise but also preserve the edge information and texture features of the image. Finally, a new method for oil spill image segmentation is proposed. Image segmentation method is a means to extract information from the image, which is the premise of image analysis. In this paper, the fuzzy C-means algorithm is improved. The histogram of the image is used as the initial value of the algorithm, which overcomes the shortcoming that the algorithm is sensitive to the initial value and can segment the oil spill part of the oil spill image more accurately.
【学位授予单位】：江苏科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP751

【共引文献】