An improved optimization method of measurement matrix for compressed sensing

The signal recovery performance of compressed sensing (CS) requires that the mutual coherence between the measurement matrix and the representing matrix should be as small as possible. In this paper an improved gradient descent method is proposed to optimize the measurement matrix. In this method, a simulated annealing (SA) learning rate factor was employed to produce the new adaptive step size and solve the antinomy between the convergence rate and accuracy. Experiment results based on Synthetic Aperture Radar (SAR) image data demonstrate that the proposed method leads to higher reconstruction performance.