基于PolInSAR的植被区高精度数字表面模型反演方法

针对传统方法无法估计电磁波在植被中的穿透深度导致植被数字表面模型(DSM)反演误差较大的问题，该文提出一种高精度DSM估计方法。该方法首先通过极化干涉相干最优中最大化相位差法分离得到电磁波在植被中高、低散射相位中心的干涉相位。然后提出一种归一化高、低散射相位中心高度随消光系数变化的模型，基于该模型搜索得到电磁波在植被中的最浅穿透深度。最后利用干涉处理方法得到高散射相位中心的高程，将最浅穿透深度补偿到该高程中，从而提升植被区DSM估计精度。利用PolSARpro软件在不同植被种类和不同植被高度下进行仿真试验以及机载全极化数据进行实测数据试验，试验结果表明该方法能有效提高植被区DSM反演精度。

High-precision Digital Surface Model Inversion Approach in Forest Region Based on PolInSAR

The general method for inversion of Digital Surface Model (DSM) in forest region has great errors due to the inestimable waves’ penetration depth. For this problem, an approach to inversion of high-precision DSM is proposed. First, the phases of high and low scattering phase centers of the waves in forest are obtained by maximizing the phase separation of the coherence optimization. Then, the normal height variation models of the high and low scattering centers with extinction factors are constructed. According to the models, the least penetration depth of the waves in forest is acquired. Eventually, by implementing the interferometric technique on the phase of high scattering phase center, a coarse DSM is retrieved, and a high-precision DSM is developed by compensating the least penetration depth to the coarse one. The validation of the method is investigated by simulated datasets of PolSARpro under different tree species and different forest heights and by airborne real datasets. It shows that the proposed method can improve the accuracy on the inversion of DSM effectively in forest region.