SATELLITE REMOTE SENSING IMAGE RESTORATION BASED ON MARKOV RANDOM FIELDS

ABSTRACT: Satellite remote sensing image is degraded by many factors such as sensor imperfections, atmospheric turbulence, electric circuit noise and satellite vibration. In this paper, a new approach for digital restoration of blurred and noisy satellite image based on markov random fields (MRF) model is presented. Maximum a posteriori (MAP) estimation framework is exploited in MRF modeling for image restoration. Under the posteriori distribution, the restoration result is derived by finding the global optimized estimation with the stimulated annealing (SA) optimization mechanism. In the experiments, the proposed method can improve both resolvable detail and contrast of blurred and noisy satellite image. The general evaluation criteria including spatial frequency, contrast, edge energy and average gradient are also adopted to quantitatively demonstrate great improvement of restoration performance by our method in comparison with the wiener filtering method. 1