Multibaseline InSAR height estimation through joint covariance matrix fitting

A new joint processing method based on joint covariance matrix fitting is presented for the multibaseline synthetic aperture radar interferometry. This method can make full use of the interferometric information embedded in the joint covariance matrix to estimate the terrain height without eigendecomposition and eigenspace division, and is insensitive to the rank variation of the joint signal subspace. By fitting the sample joint covariance matrix tapered with the coherence matrix, this method can work robustly with a finite sample support, which also makes it possible to recover the detail of terrain profile with a small neighbouring sample support. The results of numerical simulations demonstrate the validation of the proposed method.