高精度InSAR系统干涉处理算法及软件研究
发布时间：2018-07-23 12:52
【摘要】：InSAR技术是将传统的合成孔径雷达技术与干涉测量技术结合起来,由双天线模式或单天线重复轨道模式对同一地区雷达成像生成的两幅SAR复图像,利用这两幅复图像的干涉相位信息提取地表的三维信息和高程变化信息。InSAR作为一项极具研究价值的空间对地测量技术,不管是在国民经济发展领域,还是军事领域都得到了广泛的应用,其战略地位和作用引起了世界各国的高度关注。本文研究的是SAR图像后续高精度干涉处理的问题,主要是InSAR数据处理的三大关键技术：图像配准,干涉相位滤波和相位解缠。通过完善相应干涉处理算法,并编写干涉处理软件,以方便对InSAR数据进行直接处理,进一步推广InSAR技术的应用水平。主要有以下几个方面的内容：1.图像配准主要是解决空问失相关造成的失配问题。传统的图像配准算法很难同时满足配准精度和速度的要求,本文研究一种基于相关函数的分块多级图像配准方法,在保持相关函数法稳健性的基础了,对相关系数进行快速计算,大大提高了配准的速度。2.干涉相位滤波是对配准后获取的干涉相位进行相位滤波,去除噪声的影响。目前的相位滤波算法在追求滤波质量时一般会牺牲掉图像的空间分辨率或滤波速度,本文研究了一种快速干涉相位滤波方法和Sigma自适应滤波方法。前者大大提高了相位滤波的速度；后者很好地滤除了噪声并保持图像的细节特征。3.相位解缠是对滤波后的相位进行处理,恢复被模糊掉的相位周期,获取无缠绕干涉相位。针对传统的相位解缠算法不能兼顾解缠精度和速度的问题,本文研究一种基于密集残差点划分的快速相位解缠算法和基于最大堆的质量指导快速相位解缠算法。通过实际InSAR数据验证,这两种算法在相位解缠精度和速度上都得到大大的提升。4.本论文编写了InSAR系统干涉处理软件,实现了SAR图像配准、干涉相位滤波、干涉相位解缠三个干涉处理环节,最终生成了高精度无缠绕干涉相位。经过实际InSAR数据验证,本软件具有良好的稳定性,处理速度非常快,且处理精度高。
[Abstract]:InSAR technology is a combination of traditional synthetic Aperture Radar (SAR) technology and interferometric measurement technology, and two SAR complex images are generated by double antenna mode or single antenna repeat orbit mode for radar imaging in the same area. The interferometric phase information of these two complex images is used to extract 3D information and elevation change information. InSAR is regarded as a valuable spatial geodesy technique, whether in the field of national economic development. Or military field has been widely used, its strategic position and role has attracted the attention of countries all over the world. In this paper, the problem of high precision interferometry in SAR image is studied, mainly three key techniques of InSAR data processing: image registration, interferometric phase filtering and phase unwrapping. By perfecting the corresponding interference processing algorithm and compiling the interference processing software, it is convenient to process the InSAR data directly and further popularize the application level of InSAR technology. There are mainly the following aspects: 1. Image registration is mainly to solve the problem of mismatch caused by spatial correlation. The traditional image registration algorithm is very difficult to meet the requirements of registration accuracy and speed. In this paper, a block multi-level image registration method based on correlation function is studied, which maintains the robustness of the correlation function method. The correlation coefficient is calculated quickly, and the speed of registration is greatly improved. Interferometric phase filtering (IFF) is a kind of phase filtering to remove the effect of noise. The current phase filtering algorithms usually sacrifice the spatial resolution or the filtering speed of the image in pursuit of filtering quality. In this paper, a fast interferometric phase filtering method and a Sigma adaptive filtering method are studied. The former greatly improves the speed of phase filtering, while the latter filters noise well and preserves the detailed features of the image. Phase unwrapping is to process the filtered phase, recover the blurred phase period, and obtain the non-entangled interference phase. Aiming at the problem that the traditional phase unwrapping algorithm can not take into account the unwrapping accuracy and speed, a fast phase unwrapping algorithm based on dense residual handicap partition and a mass-guided fast phase unwrapping algorithm based on maximum heap are studied in this paper. Through the actual InSAR data validation, the two algorithms are greatly improved in the accuracy and speed of phase unwrapping. In this paper, the interference processing software of InSAR system is written, which realizes three steps of SAR image registration, interferometric phase filtering and interference phase unwrapping. Finally, the high-precision non-winding interferometric phase is generated. The software has good stability, fast processing speed and high processing precision.
【学位授予单位】：长沙理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN957.52
，
本文编号：2139472
[Abstract]:InSAR technology is a combination of traditional synthetic Aperture Radar (SAR) technology and interferometric measurement technology, and two SAR complex images are generated by double antenna mode or single antenna repeat orbit mode for radar imaging in the same area. The interferometric phase information of these two complex images is used to extract 3D information and elevation change information. InSAR is regarded as a valuable spatial geodesy technique, whether in the field of national economic development. Or military field has been widely used, its strategic position and role has attracted the attention of countries all over the world. In this paper, the problem of high precision interferometry in SAR image is studied, mainly three key techniques of InSAR data processing: image registration, interferometric phase filtering and phase unwrapping. By perfecting the corresponding interference processing algorithm and compiling the interference processing software, it is convenient to process the InSAR data directly and further popularize the application level of InSAR technology. There are mainly the following aspects: 1. Image registration is mainly to solve the problem of mismatch caused by spatial correlation. The traditional image registration algorithm is very difficult to meet the requirements of registration accuracy and speed. In this paper, a block multi-level image registration method based on correlation function is studied, which maintains the robustness of the correlation function method. The correlation coefficient is calculated quickly, and the speed of registration is greatly improved. Interferometric phase filtering (IFF) is a kind of phase filtering to remove the effect of noise. The current phase filtering algorithms usually sacrifice the spatial resolution or the filtering speed of the image in pursuit of filtering quality. In this paper, a fast interferometric phase filtering method and a Sigma adaptive filtering method are studied. The former greatly improves the speed of phase filtering, while the latter filters noise well and preserves the detailed features of the image. Phase unwrapping is to process the filtered phase, recover the blurred phase period, and obtain the non-entangled interference phase. Aiming at the problem that the traditional phase unwrapping algorithm can not take into account the unwrapping accuracy and speed, a fast phase unwrapping algorithm based on dense residual handicap partition and a mass-guided fast phase unwrapping algorithm based on maximum heap are studied in this paper. Through the actual InSAR data validation, the two algorithms are greatly improved in the accuracy and speed of phase unwrapping. In this paper, the interference processing software of InSAR system is written, which realizes three steps of SAR image registration, interferometric phase filtering and interference phase unwrapping. Finally, the high-precision non-winding interferometric phase is generated. The software has good stability, fast processing speed and high processing precision.
【学位授予单位】：长沙理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN957.52
，
本文编号：2139472
本文链接：https://www.wllwen.com/kejilunwen/wltx/2139472.html
