一种基于多普勒谱AR建模的SAR图像纹斑噪声抑制技术

Despeckling of SAR Images by AR Modeling of Doppler Spectrum

基于合成孔径雷达(SAR)成像与自回归(AR)模型的特点,提出了一种SAR图像纹斑噪声抑制技术。将SAR复值图像的多普勒谱表示成一定阶数的AR模型,根据合适的AR模型系数估算SAR强度或幅值图像可保持较高的分辨率。给出了方法的模型和步骤。实验结果表明:所获幅值图像的纹斑噪声抑制,以及边缘和纹理特征等细节保持都优于常用的基于SAR图像多普勒谱的纹斑噪声抑制——多视处理技术。

Based on the characteristics of synthetic aperture radar(SAR) imaging and auto-regressive(AR) model,a new technique to despeckle SAR images was put forward in this paper.The Doppler spectrum was expressed as an auto-regressive model with a certain order.The intensity or amplitude image of SAR acquired under appropriate coefficients AR model had a high resolution.The test results showed that the despeckling and detail keeping of the edge and grain of the amplitude image gained by this method were better than...