Wishart-H/Alpha法进行聚类可以实现复杂场景的高精度分类。然而在聚类的过程中,由于各类中心所对应的散射机理发生了混淆,使得不透水层代表之一的水泥路面与裸露土壤发生了混淆,这对实际应用不利。提出了利用最优相干系数与极化总功率系数构成的二维直方图空间进行阈值分割,将Wishart-H/Al-pha分类方法中混淆的水泥道路与裸露土壤重新分离出来,并通过国内机载X波段双天线极化干涉实验,验证了本方法的有效性。
It's suitable to distinguish the scattering properties of different objects basing on Wishart-H/Alpha classification.However,various scattering mechanisms are merged when class clustering,which confuse the road and the bare soil.In this paper,an improved scheme is performed which segments the optimal-coherence and Polarimetric-Span 2D-histogram to subdivide the road and bare soil after Wishart-H/Alpha classification.Finally,the experiment on domestic dual-antenna X band POLINSAR datum proves that proposed method can separate roads and bare soil effectively.