Interferometric Synthetic Aperture Radar (InSAR) has been proposed as a technique for the reconstruction of the Digital Elevation Model (DEM),which is widely used in global terrain mapping mission [1].The DEM reconstruction methods of InSAR technique are usually based on the so-called phase unwrapping (PU) operation.In highly sloped terrain,the interferometric phase gradients (IPG) of the adjacent pixels are more than 7π.Therefore,the traditional PU methods cannot correctly obtain the absolute interferometric phase (IP) [2].Several methods based on multifrequency InSAR (MF-InSAR),such as the maximum likelihood (ML) method,maximum a posteriori (MAP) method,and so on,have been widely studied.The ML method can quickly obtain the estimation of target heights [3].However,its performance largely depends on the number of interferograms,and the estimation results also can be easily influenced by the phase noise.Compared to the ML method,the MAP method can reduce the number of interferograms,but it suffers from some other limitations,such as the computational time [4].