基于改进模拟退火遗传算法的 INSAR 相位解缠算法

摘要

为了解决经典的 Goldstein 枝切线法容易生成过长的枝切线和较多封闭区域的问题，提出一种基于改进模拟退火遗传算法的 INSAR（Interferometric Synthetic Aperture Radar）相位解缠算法。该算法首先对部分残差点进行预处理，生成极性平衡的小段枝切线，然后使用改进模拟退火遗传算法求解剩余残差点的优化组合。经这两步处理后，所得到的枝切线的总长度和封闭区域的数量都明显减少。对真实 INSAR 数据的实验结果表明，该算法在运行时间和解缠精度上均有一定的优越性。%In order to solve the problems that classical Goldstein’s branch-cuts method easily generates excessively long branch-cuts and more enclosed areas,we proposed a phase unwrapping algorithm for interferometric SAR,which is based on improved stimulated annealing genetic algorithm.First,the algorithm pre-processes part of residues to generate small piece branch-cuts with balanced polarity.Then it uses improved stimulated annealing genetic algorithm to calculate the optimised combination of remaining residues.After these two processing steps,the total length of branch-cuts derived and the number of enclosed areas decrease significantly.Results of experiment on real INSAR data proved that the proposed algorithm has certain advantage in run time and phase unwrapping precision.