融合CZT频谱细化算法的InSAR相位滤波方法

InSAR Phase Filtering Method Based on CZT Spectrum Thinning Algorithm

为了有效抑制噪声并更好保持干涉条纹细节信息,使得干涉相位滤波的结果有利于相位解缠和提高干涉测量精度,提出了一种融合频谱细化的InSAR自适应相位滤波方法。将干涉相位进行傅里叶变换到频率域,利用线性调频Z变换(CZT)频谱细化的方法,对干涉相位频域频谱进行估计、细化,提高对干涉相位噪声的分辨能力,并结合自适应滤波的方法以提高干涉相位的滤波效果和更好保持相位细节信息的特点。将该方法与Boxcar滤波、非局部均值滤波以及自适应Goldstein滤波方法做对比,在仿真实验与实测实验中使用该方法滤波后产生的相位残差点数量最少,分别相较于滤波前减少了99.4%与92.8%,在相位导数标准差及仿真实验误差上较其他3种方法均有不同程度的提升,证明了该方法的有效性。

This paper proposes an InSAR adaptive phase filtering method with spectrum thinning to effectively suppress noise and preserve the details of the interference fringes, so that the results of interference phase filtering are beneficial to phase unwrapping and accuracy improvement of interferometry. The interference phase is converted to the frequency domain using the Fourier transform, and the frequency domain spectrum of the interference phase is estimated and refined using the frequency spectrum thinning method of the Chirp-Z Transform(CZT),so as to enhance the resolution capability of interference phase noise. Combined with the adaptive filtering method, the interference phase filtering effect is enhanced and the phase information retention characteristics are enhanced. Comparing this method with boxcar filter, non-local means filter, and adaptive Goldstein filter, the number of phase residuals generated by this method is reduced by 99.4% and 92.8%,respectively, compared with those before filtering. Compared with the other three methods, the standard deviation of phase derivative and simulation error are improved to some extent, which proves the effectiveness of this method.