基于多重网格方法的条纹图像相位解包技术

摘要光学层析技术所获取的投影数据不同于CT技术中简单的一维数据，它是某种被调制的二维信号。因此，投影数据的提取是光学层析技术的重要一步。傅里叶变换提取相位克服了条纹跟踪法抗噪能力低、不能获取整个条纹信息以及相移法对采集机械要求过高的缺点，是较为优良的相位提取方法43108。

本文为了克服傅里叶变换中手动提取频谱不理想的问题，采用了自动提取频谱的方法。这个方法可以准确的选取频谱的中心及旁瓣，并对不同的条纹图保持一致的频谱选择标准，避免人为选择时的主观性干扰。

但由于图像噪声的存在，在条纹图像的解包过程中不能依赖理想的解包路径进行，必须根据噪声特性自动选择解包路径。多重网格方法相较于其他解包方法有较强的抗噪能力，能够精确的解包相位。本文就是通过多重网格方法来研究光学层析的投影条纹图像处理技术中的相位解包问题。

毕业论文关键词: 光学层析，相位提取，相位解包，多重网格

毕业设计说明书外文摘要

Title Based on the multiple stripe image phase to unpack the technology of grid method

Abstract

Optical tomographic techniques acquired projection data is different from the simple one dimensional data in CT technology, it's kind of a two-dimensional modulated signal. Therefore, the projection data extraction is an important step in optical tomography technology. Fourier transform to extract phase to overcome the tracking method of stripe anti-noise property is low, can't get the entire stripe information as well as the shortcoming of phase shift method to obtain high mechanical requirements, is relatively good phase extraction method.

Manual extraction in this paper in order to overcome the Fourier transform spectrum is not ideal, adopted the approach of automatic extraction spectrum. This method can accurately select the center of the frequency spectrum and the side lobe, and the fringe pattern to maintain consistent spectrum selection criteria, to avoid the subjectiveness of human choice.

But because of the existence of image noise in the process of stripe images to unpack the package cannot rely on the ideal solution path, must choose according to the noise characteristics of automatic unpacking path. Multigrid method is compared with other unpacking method has strong anti-noise property, can accurate solution phase. This article is through multiple grid method to study the optical tomographic projection fringe phase unpacking problem in image processing.

Keywords: Optical tomography ， Phase extraction ，phase unwrapping ，Multiple grid

1 引言1

1.1计算机层析技术1

1.2光学层析技术1

1.3投影相位提取技术1

1.3.1条纹跟踪法2

1.3.2相移法2

1.3.3傅里叶变换法2

1.4相位解包技术3

1.5 本文内容安排 4

2 傅里叶变换方法提取包裹相位信息5

2.1傅里叶变换提取相位原理5

2.2相位场模拟5

2.3傅里叶变换方法提取包裹相位信息模拟6

2.4本章小结10

3 相位解包方法的研究11

3.1相位解包原理11

3.2.1 基本迭代法11

3.2.2 Gauss-Seidel迭代法12

3.2.3 Gauss-Seidel迭代解包算法13

3.3 多重网格解包算法14

3.4 解包算法的比较17

3.5 本章小结18

4 多方向莫尔实验19

4.1多方向投影相位装置19

4.2莫尔条纹图的相位提取20

结论 27

致谢 28

参考文献29

1 引言

1.1 计算机层析技术基于多重网格方法的条纹图像相位解包技术:http://www.751com.cn/wuli/lunwen_43788.html