论文部分内容阅读
目的相位解缠是In SAR干涉数据处理的关键步骤,而解缠不连续(即相位跳变)问题却普遍存在,尤其在机载In SAR系统中,由于数据的高分辨率,使得低矮地物如树木带在数据中表现为相位不一致,因而相位跳变问题更加显著。星载In SAR相位解缠广泛使用统计费用网络流(SNAPHU)算法~([1]),借鉴其经验将SNAPHU算法引入高分辨机载In SAR相位解缠。而残差点退化方法能有效补偿局部相位不一致区域。因此本文提出一种结合残差点退化方法与SNAPHU算法的高分辨率机载In SAR相位解缠算法。方法将原始In SAR数据滤波且去除平地相位,再对其进行残差点退化处理。残差点退化包含残差点定位,及残差点补偿两部分。根据残差点及其邻域像元的性质,对残差点进行补偿使其退化为非残差点,不断迭代这一过程,以减少图像中的残差点,优化局部数据。根据机载In SAR系统定标参数,修正SNAPHU算法中的参数及几何模型,使用修正后算法进行相位解缠。结果利用2011年四川江油地区的单轨双天线X波段机载In SAR数据进行了试验,试验结果表明,在相位不一致,相干性低的连续树木带区域,该算法显著缩小了解缠相位不连续区域,修正了大面积的相位跳变。结论验证了残差点退化方法结合统计费用网络流算法可有效解决解缠相位大面积跳变问题,且对噪声具有鲁棒性。
The purpose of phase unwrapping is the key step of In SAR interferometric data processing. However, the problem of unwrapping discontinuity (ie phase jump) is ubiquitous. Especially in airborne In SAR system, due to the high resolution of data, Objects such as trees and bands show inconsistent phases in the data, and the problem of phase jumps is even more pronounced. SNAPHU algorithm ([1]) is widely used in the phase-unwrapping of In-Plane In SAR. The SNAPHU algorithm is introduced into phase-unwrapping of high-resolution airborne In SAR based on its experience. The residual point degradation method can effectively compensate for the local phase inconsistent region. Therefore, this paper presents a high resolution in-plane In-SAR phase unwrapping algorithm combining the method of residual point degeneration and SNAPHU algorithm. The method filters the original In SAR data and removes the flat phase, then carries on the residual point degeneration processing. Residual point degradation includes residual point location and residual point compensation in two parts. According to the properties of residual pixels and their adjacent pixels, the residuals are compensated to degenerate into non-residuals, and the process is iterated continuously to reduce the residuals in the image and optimize the local data. According to the calibration parameters of airborne In SAR system, the parameters and geometric model in SNAPHU algorithm are modified and the phase unwrapping algorithm is used. Results The X-band airborne In SAR data of monorail and double antenna in Sichuan Jiangyou area in 2011 were used in the experiment. The experimental results show that this method significantly reduces the unwinding phase discontinuities in the continuous tree belt with inconsistent phase and low coherence, Fixed a large area of phase transition. The conclusion verifies that the residual point degradation method combined with the statistical cost network flow algorithm can effectively solve the large-area jump problem of the unwrapping phase and is robust to noise.