高精度InSAR配准及相位解缠技术研究

发布时间：2018-05-14 09:17

本文选题：高精度InSAR + 配准　；参考：《电子科技大学》2014年硕士论文

【摘要】：具有全天候、全天时测量能力的合成孔径雷达干涉(InSAR)测量技术,在大范围地形高程测绘方面具有广泛的应用前景。本文针对高精度InSAR配准和相位解缠技术进行了详细研究,分析了相关算法优缺点,并提出了改进的算法,最后结合仿真及实测数据对比分析了各算法性能。所做的具体工作及创新如下:(1)研究了基于DFT模型的高效精配准算法。本文指出了InSAR主副图像地距分辨率不均匀导致该算法达不到配准要求的问题,提出了使用分块策略进行精配准的方法,使用实测数据进行了验证,针对配准结果只有部分干涉纹的情况,分块处理在大部分干涉相位图都能得到干涉纹,大大提高了配准效果。(2)研究了路径跟踪类相位解缠算法。首先分析了Goldstein枝切法的优缺点。研究并指出了Goldstein枝切法布置枝切线时的不足,在此基础上提出了改进的枝切法。改进的枝切法和Goldstein枝切法相比能获得更好的枝切线,解缠结果误差少,且消除了Goldstein枝切法解缠结果可能存在的解缠空洞,明显提高了解缠效果。最后本文仿真了不连续相位曲面的解缠情况,指出了该类算法的适用条件,同时也说明了对存在不连续时实测数据解缠处理该类算法可能失败。(3)研究了最小范数类相位解缠算法。首先指出了最小二乘法相位解缠算法和迭代最小二乘相位解缠算法解缠结果是真实相位的估计值,解缠误差较大,分布广泛。在此基础上提出改进的迭代最小二乘相位解缠算法,通过推导公式发现,当解缠相位与真实相位的差不存在缠绕时,可以将解缠相位与缠绕相位的相位差主值加入到解缠相位中,得到精确的解缠相位。改进算法对仿真和实测数据解缠只有非常少的解缠误差,验证了改进算法的解缠性能。同时本文也通过仿真不连续相位面,说明了该算法不适用于不连续相位解缠。最后研究了图割相位解缠算法,该算法通过将相位解缠问题转化为图像分割问题,每次对分割得到的两部分分别加0或者2?,迭代到不能分割得到解缠相位。该算法能够识别不连续,打破连续性条件限制。同时其解缠精度非常高,仿真和实测数据都证明了该算法的有效性。总之,本文提出了InSAR高精度图像配准算法和相位解缠算法,为高精度InSAR数据处理提供了新的研究思路。
[Abstract]:Synthetic aperture radar interferometry (InSAR) measurement technology with all-weather and all-weather measurement is widely used in large range terrain elevation surveying and mapping. In this paper, the high precision InSAR registration and phase unwrapping technology are studied in detail, the advantages and disadvantages of the related algorithms are analyzed, and the improved algorithm is put forward. Finally, the simulation is combined with imitation. The performance of each algorithm is compared and analyzed. The specific work and innovation are as follows: (1) the efficient and accurate registration algorithm based on the DFT model is studied. This paper points out the problem that the uneven resolution of the InSAR main and vice image leads to the problem that the algorithm can not reach the registration requirements, and proposes a method of using the block strategy to carry out the precise registration. The measured data are verified. In case the registration results only partially interfere with the pattern, the block processing can get the interference pattern in most interference phase diagrams, which greatly improves the registration effect. (2) the phase unwrapping algorithm of path tracking class is studied. First, the advantages and disadvantages of the Goldstein branch cutting method are analyzed. The Goldstein char method is studied and pointed out. The improved branch cutting method is proposed on this basis. The improved branch cutting method can obtain better branch tangent compared with the Goldstein branch cutting method, and the error of the unwrapping result is less, and the unwrapping holes that may exist in the unwrapping result of the Goldstein branch cutting method are eliminated, and the entangling effect is obviously improved. Finally, the discontinuous phase is simulated. The application conditions of this kind of algorithm are pointed out. At the same time, the possible failure of this kind of algorithm is explained. (3) the minimum norm phase unwrapping algorithm is studied. First, it is pointed out that the least squares phase unwrapping algorithm and the iterative least squares phase unwrapping algorithm are true. An improved iterative least square phase unwrapping algorithm is proposed. By deriving the formula, it is found that when the difference between the unwrapping phase and the real phase is not entangled, the phase difference between the unwrapping phase and the entangled phase can be added to the unwrapping phase, and the exact solution can be obtained. The improved algorithm has only very little unwrapping error for the simulation and the measured data, and verifies the unwrapping performance of the improved algorithm. At the same time, this paper also shows that the algorithm is not suitable for discontinuous phase unwrapping by simulation of the discontinuous phase surface. Finally, the graph cut phase unwrapping algorithm is studied. The algorithm transfers the phase unwrapping problem by turning the phase unwrapping problem. The algorithm is transformed into image segmentation problem. Each time the two parts of the segmentation are added to 0 or 2 respectively, the unwrapping phase can not be segmented. The algorithm can recognize discontinuity and break the continuity condition limit. At the same time, its unwrapping accuracy is very high. Both the simulation and the measured data prove the effectiveness of the algorithm. In a word, the high precision of InSAR is proposed. Image registration algorithm and phase unwrapping algorithm provide a new research idea for high precision InSAR data processing.

【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN957.52

【参考文献】