REMOTE SENSING IMAGE RESTORATION FOR ENVIRONMENTAL APPLICATIONS USING ESTIMATED PARAMETERS

Image restoration is the process of cleaning of distorted image and reinstating to its original form. In remote sensing images, usually the degradation of images happens during the acquisition process of images mostly affected by blur. This paper proposes a novel remote sensing image restoration technique. The technique mainly concentrates on the parameter estimation which is used to estimate the PSF (Point Spread Function). First, the image is transformed to the frequency using Fourier transform. Then, Gaussian Low-pass filter is used to remove or eliminate some noises present in the corrupted image. Next, the PSF or degradation function is estimated using parameters estimation based on blur angle and blur length. For effectiveness and simplicity, Radon transformation was used to estimate these parameters, which later were used to estimate the PSF of image. From the estimated PSF the image is restored using Wiener filter. In order to improve the quality of restored image, a post-processing process is added to enhance the image. The evaluation results shows that the proposed technique is effective and better compared to the existing techniques such as Lucy-Richardson and RSL Adaptive filters.