Spotlight SAR images restoration based on tomography model

Abstract

A novel approach for spotlight synthetic aperture radar (SAR) image restoration based on tomography model is presented for the characteristic of spotlight SAR image. Giving an initial guess value as the external incentives to detect SAR images of the internal features, the forward model is established to obtain the theoretical value for imaging. In order to overcome ill-posed problem of restoration and remain more texture information of targets in SAR image, adaptive regularization parameters are introduced, which make it possible to run synchronously with iteration operation and automatically modify the model parameters to the optimal value. Meanwhile, according to forward model, inversion model of SAR image is determined by using adaptive modified matrix. The updating value of image restoration is obtained by solving equations and setting the threshold as a condition for the termination of iteration so that the final results of the image restoration are obtained. The experimental results show that the proposed algorithm is effective for SAR image restoration with blurring and noises.