Image denoising is a key issue in all image processing researches. The great challenge of image denoising is how to preserve the edges and all fine details of an image when reducing the noise. In this paper, a comparative study of image denoising techniques for underwater SONAR (Sound Navigation and Ranging) images relying on spatial filters is presented. In particular four types of spatial filters (Average, Gaussian, Laplacian of Gaussian and Median filters) are applied to judge the

more »

... On each image, different window size configurations starting from 3x3 to 29x29 are applied and the performances of image filtering techniques are analyzed by the estimation of parametric values such as Mean Square Error (MSE), Peak Signal to Noise Ratio (PSNR) and execution time for filtering the images. It is observed that by increasing the window size the execution time will increase and PSNR values will decrease. With this analysis and from the results it is found that the optimum filter is Gaussian and the optimum window size is 3×3 for the underwater SONAR images which gives the best execution time, MSE and the PSNR value (37.87dB). KEYWORDS: Denoising, SONAR, Mean Square Error (MSE), Peak Signal to Noise Ratio (PSNR) INTRODUCTION The term spatial domain refers to the image plane itself, and the methods in this category are based on direct manipulation of pixels in an image. In this paper the main focus is on two important categories in spatial filtering i.e. linear and nonlinear spatial filtering. The spatial filtering is also called as neighbourhood processing. An image can be modified by applying a particular function to each pixel value. The neighborhood processing may be considered as an extension of this, where a function is applied to a neighborhood of each pixel. The idea is to move a "mask (kernel)": a rectangle (usually with sides of odd length) or other shape over the given image. Depending on the computations performed on the pixels of neighborhoods the operations is called as linear or nonlinear spatial filtering and are clearly described in this paper. Though various filters are already available in the open literature, here this paper focuses on finding out an optimum filtering technique for underwater SONAR images. For this purpose, several spatial filtering techniques such as Average, Gaussian, Laplacian of Gaussian and Median filters are analyzed by the estimation of parametric values such as Mean Square Error (MSE), Peak Signal to Noise Ratio (PSNR) and execution time for filtering the images. SPATIAL FILTERS Linear spatial filtering: If the function by which the new gray value is calculated is a linear function of all the grey values in the mask, then the filter is called linear filter. Example of a linear filter is average filter. A linear filter can be implemented by convolving the mask with the input image. For a 3x5 mask, the convolution is shown in Eq. 1: