不同质量图在卡尔曼滤波相位解缠中应用探讨
The Study on Application of Different Quality Maps in Kalman Filter Phase Unwrapping

卡尔曼滤波将相位解缠转化为状态估计问题，实现相位解缠与噪声消除的一并处理。通过建立相位的动态方程和观测方程来求解真实相位，在常规方法中观测方程的噪声方差由相干图的值来确定。本文采用三种不同质量图的值来确定噪声方差，实施扩展卡尔曼滤波相位解缠算法。分别在地形平坦和陡峭两种条件下对InSAR数据进行实验，通过对解缠结果进行对比分析，表明对于地形陡峭数据应用相位导数方差图得到的结果较可靠，而地形平坦数据应用伪相干图能够得到较为精确的解缠结果。
The Kalman Filter transforms the phase unwrapping problem into state estimate issue to deal with phase unwrapping and noise eliminating at the same time. The true phase is obtained through establishment of the dynamic equation and observation equation, noise variance of observation equation is determined by the value of coherence image in the original method. In this paper, noise variance is determined by three kinds of different quality maps, then the Extended Kalman Filter phase unwrapping is implemented, using InSAR data to do the experiment in the steep and flat terrain. Through comparing and analyzing the phase unwrapping results, it is verified that the result using phase derivative variance map is more reliable in the steep terrain, but in the flat terrain, the result using pseudo-coherence map is more precise.