超高分辨率PolSAR图像去噪对极化特征的影响

Influence of De-speckling on Polarimetric Characteristics for Very High Resolution Polarimetric SAR Images

随着超高分辨率极化合成孔径雷达(PolSAR)图像数据的出现,有必要对PolSAR图像的去噪过程对于极化分解结果的影响进行评测。相干斑去噪过程的实施将对观测场景中的PolSAR图像原有极化属性产生不同程度的影响,这将会导致分解结果出现偏差。以非相干分解模型为评估对象,以喷气推进实验室(JPL)提供的开源机载超高分辨率PolSAR系统数据为实验样本,提出了一种利用绝对相对误差的定量评测方法。评测结果显示:对于超高分辨率图像,相干斑噪声的影响明显变小,分解窗口可相应的缩小为7×7的尺寸。研究结果对于PolSAR图像的后期遥感应用中的参数选择具有重要参考意义。

With the popularization of the very high resolution polarimetric synthetic aperture radar（PolSAR） images,it is necessary to make an evaluation of the influence of the speckle filtering on the decomposition.After the de-speckling processing,the inherent polarimetric properties of the SAR imaging data may be more or less altered.Even a slight change may generate significant bias of the decomposition.The incoherent decomposition is employed for evaluation,and then the scene data with very high resolution provided by the Jet Propulsion Laboratory（JPL） is utilized for the experiment.An evaluating method based on absolute relative bias is introduced subsequently.The corresponding results demonstrate that the impact of speckle becomes weak and the decomposition window can be reduced to the size of 7×7.This quantitative measurement research will be beneficial to the parameter selection for follow-up remote-sensing applications based on PolSAR images.