Abstract: Phase unwrapping of Interferometric Synthetic Aperture Radar（InSAR） is one of the most important steps to obtain an accurate DEM（Digital Elevating Model） and reconstruct 3-D landform images.The theories of interferometric and InSAR 2-D phase unwrapping based on least square are presented.Unweighted multi-grid algorithm which is used to solve Partial Differential Equation（PDE） on larger grids with fast calculating rate is proposed to solve the discrete Poisson equation.The method can transform low frequency components to high frequency components，wipe off quickly using Gauss-Seidel relaxation and smoothen multi-frequency errors.It is indicated that the convergence speed is the same as FFT and DCT with higher unwrapping accuracy.

摘要： 相位解缠是干涉合成孔径雷达（InSAR）测量地形高程、重建三维地形模型的重要步骤之一。介绍了最小二乘的相位解缠原理，提出了一种基于无权重的多重网格算法求解离散泊松方程。该算法是在较大网格基础上解决偏微分方程（PDE）的快速方法，将误差的低频成分转换为高频成分，这样就可以使用高斯-塞德尔松弛法去除多种频率的误差分量。实验结果证明：该方法具有收敛速度快，解缠精度高等优点。