首页|粒子-平滑变结构滤波相位解缠算法

粒子-平滑变结构滤波相位解缠算法

扫码查看
为了有效提升干涉图相位解缠精度,提出粒子-平滑变结构滤波相位解缠算法:①把融入残差因子的平滑变结构滤波算法应用于干涉图相位解缠,结合局部相位梯度估计技术,构建粒子-平滑变结构滤波相位解缠程序;②把干涉图相位微分偏差图作为质量图,嵌入堆排序路径跟踪策略指导相位解缠路径;③利用归一化的干涉图相位微分偏差图把干涉图划分为3个不同质量层级的缠绕像元集,粒子-平滑变结构滤波相位解缠程序沿第1质量层级→第2质量层级→第3质量层级的顺序依次解缠干涉图缠绕像元,获得干涉图所有缠绕像元的解缠相位值.模拟和实测数据实验结果表明与一些代表性算法相比,上述方法能较好地执行干涉图相位解缠任务,显现了较好的鲁棒性,有效提升了干涉图相位解缠精度.
Particle-smooth variable structure filtering phase unwrapping algorithm
To improve the accuracy of the phase unwrapping for interferogram,a particle-smooth variable structure filtering phase unwrapping algorithm is presented. First,the smooth variable structure filter (SVSF )incorporating residual factors is applied to phase unwrapping for interferogram,and is combined with an efficient local phase gradient estimation technique to construct a particle-smooth variable structure filtering phase unwrapping(PF-SVSFPU)procedure. Second,phase derivative variance map of the interferogram is used as a quality map and embedded into a heapsort-based path-following strategy to guide the working paths of the PF-SVSFPU procedure;Finally,the normalized phase derivative variance map of the interferogram is used to divide the interferogram into three groups of wrapped pixels with different quality levels,and the PF-SVSFPU procedure sequentially unwraps the wrapped pixels along the order of the first quality level group→the second quality level group→the third quality level group to obtain the unwrapped phase of all wrapped pixels of the interferogram. Results delivered by the experiments on synthetic and measured data show that the proposed method can perform phase unwrapping task for interferograms better compared to some representative methods,demonstrating good robustness and effectively improving the accuracy of phase unwrapping for interferograms.

phase unwrappingparticle filtersmooth variable structure filtercombined filter

李蓉、谢先明

展开 >

广西科技大学 自动化学院,广西 柳州 545006

广西科技大学 电子工程学院,广西 柳州 545006

相位解缠 粒子滤波 平滑变结构滤波 组合滤波

国家自然科学基金项目广西自然科学基金项目

621610032023GXNSFAA026024

2024

测绘科学
中国测绘科学研究院

测绘科学

CSTPCD北大核心
影响因子:0.774
ISSN:1009-2307
年,卷(期):2024.49(1)
  • 28