残差点退化的统计费用网络流机载相位解缠算法

发布时间：2018-10-11 07:59

【摘要】：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。
[Abstract]:Objective Phase unwrapping is a key step in In SAR interferometric data processing, but unwrapping discontinuity (i.e. phase jump) is a common problem, especially in airborne In SAR systems, because of the high resolution of data. Therefore, the phase jump problem of low ground objects such as tree belt is more obvious because of the inconsistent phase in the data. Spaceborne In SAR phase unwrapping is widely used in statistical cost network flow (SNAPHU) algorithm. The SNAPHU algorithm is introduced to high-resolution airborne In SAR phase unwrapping based on its experience. The residual handicap degradation method can effectively compensate the local phase inconsistency region. In this paper, a phase unwrapping algorithm for high resolution airborne In SAR is proposed, which combines residual handicap degradation method and SNAPHU algorithm. Methods the original In SAR data were filtered and the flat phase was removed, and then the residual handicap degradation was carried out. Handicap degeneration includes handicap location and handicap compensation. According to the properties of the residual handicap and its adjacent pixels, the handicap is compensated to degenerate into non-handicap, and this process is iterated continuously in order to reduce the residual handicap in the image and optimize the local data. According to the calibration parameters of airborne In SAR system, the parameters and geometric models of the SNAPHU algorithm are modified, and the modified algorithm is used to unwrap the phase. Results based on the airborne In SAR data of single track and dual antenna in 2011, the experimental results show that in the continuous tree-belt region, where the phase is inconsistent and the coherence is low, The algorithm reduces the region of phase discontinuity and corrects the large area of phase jump. Conclusion it is proved that the residual handicap degradation method combined with the statistical cost network flow algorithm can effectively solve the large area jump problem of unwrapping phase and is robust to noise.
【作者单位】：中国科学院遥感与数字地球研究所;三亚中科遥感所;
【基金】：海南省重点研发计划基金项目(SY16ZY02132)~~
【分类号】：TN957.52

【相似文献】