Blind restoration of remote sensing images
Sakshi Mittal, Amit Garg · 2013
Images invariably represent a blended version of the native scene. Restoration of remote sensing images from multiple degraded images taken under diverse happenstances is the major concern of this work. The work is concern with blind restoration of remote sensing images which symbolizes restoring the images without knowing their primitive extent and the degradation cause. Existing multi channel blind restoration methods require spatially aligned images and correct veneration of blur size and shape. A brief study of degradation phenomenon along with heterogeneous restoration techniques is also presented. The proposed method is a small modification of existing LPA-ICI algorithm by using a different wavelet technique (Immerkaer-Daubechies). Simulation results are presented for standard and real grayscale and colored images. Subjective and Objective evaluation of the restored images is presented for quality assessment. From the results presented it can be concluded that for all the images a high value of Peak Signal to Noise Ratio (PSNR) is obtained by using proposed algorithm as compared to LPA-ICI.