Spatial resolution enhancement of sonar images using one step pixel prediction

In this paper, a one step pixel prediction based approach for sonar image estimation in speckle noise is proposed. Spatial resolution of sonar images are degraded by the multiplicative noise namely speckle. Denoising is required in sonar images to characterize different region by analyzing the image. First the homomorphic approach is applied on the reflectivity domain relation of the sonar image to obtain the observation model for the filter. Kalman estimation is then used to estimate the original pixel value from its degraded observations. Simulated results are provided to validate the proposed method.