空间遥感图像退化机理及复原关键技术研究

发布时间：2018-06-22 05:16

本文选题：图像复原 + 调制传递函数　；参考：《长春理工大学》2014年硕士论文

【摘要】：空间遥感图像的形成经历了多个环节,这些环节中不可避免的出现了图像的退化,图像质量降低,为了获得高质量的空间图像,需要对获得的遥感图像进行消噪和去模糊处理,一般进行图像增强和图像复原操作。目前国内外各个机构对获得高质量的遥感图像做了研究,因此而衍生出了众多的图像恢复方法,例如采用图像点扩散函数的图像复原,采用图像调制传递函数的图像复原。本文深入研究了利用刃边法求取点扩散函数的过程,对影响图像复原效果的因素进行了分析,并提出了自己的评价体系用于遥感图像的复原效果评价。本文的主要工作如下： 1、针对图像求取点扩散函数时的数据拟合时的精确度问题,本文提出了一种基于评价效果的最优费米函数拟合方案用于图像边缘扩散函数的拟合。 2、对各个复原算法进行了复原效果的验证,并深入探究了Lucy-Richardson复原方法中复原次数对遥感图像的复原效果的影响。 3、针对图像的评价效果,提出了一套基于复原后图像噪声的图像评价方案用于复原后图像的评价,实验证明了该方案的有效性。 4、深入研究各个评价算法,针对图像评价效率的问题,提出了基于选定区域的评价效果代替整体图像评价效果的方案,该方案在实验中也得到了有效性验证。 5、总结归纳当前的图像评价方法,利用图像评价中的高效算法,结合数学建模中的求取极值的方法,提出了一套评价体系用于复原后图像的评价。实验证明了该方案针对遥感图像评价的有效性。 6、采用vC与Matlab混合编程进行了实验流程的验证,采用VC多线程编程验证了图像的复原效果。
[Abstract]:The formation of space remote sensing image has experienced many links. In these links, the degradation of the image is inevitable, and the image quality is reduced. In order to obtain the high quality space image, it is necessary to de-noising and deblurring the obtained remote sensing image. General image enhancement and image restoration operations. At present, various institutions at home and abroad have done research on obtaining high quality remote sensing images, so many image restoration methods have been derived, such as image restoration using image point diffusion function, image restoration using image modulation transfer function. In this paper, the process of obtaining point diffusion function by edge method is studied, and the factors influencing the restoration effect are analyzed, and the evaluation system is put forward to evaluate the restoration effect of remote sensing image. The main work of this paper is as follows: 1. Aiming at the accuracy of the data fitting of the point diffusion function, In this paper, an optimal Fermi function fitting scheme based on evaluation effect is proposed for image edge diffusion function fitting. 2. The restoration effect of each restoration algorithm is verified. The influence of restoration times on the restoration effect of remote sensing image in Lucy-Richardson restoration method is discussed. 3. In view of the image evaluation effect, an image evaluation scheme based on reconstructed image noise is proposed to evaluate the reconstructed image. The experiment proves the effectiveness of the scheme. 4. The evaluation algorithm is studied deeply. Aiming at the problem of image evaluation efficiency, the evaluation effect based on the selected region is proposed to replace the overall image evaluation effect. The scheme is also validated in the experiment. 5. The current image evaluation method is summarized, and the efficient algorithm in image evaluation is used, combined with the method of calculating extreme value in mathematical modeling. An evaluation system is proposed for image evaluation after restoration. The experimental results show that the scheme is effective for remote sensing image evaluation. 6. The experiment flow is verified by mixed programming of VC and Matlab, and the effect of image restoration is verified by VC multithreading programming.
【学位授予单位】：长春理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP751

【参考文献】