DESIGN AND IMPLEMENTATION OF TWO STEP FUZZY FILTER FOR IMAGE PROCESSING

Abstract

Noise reduction from images is an important step in many image-processing applications. In the presence of noise, filtering of noisy pixel and its correction is an important pre-processing step if they are to be followed by other tasks such as edge detection, feature extraction and object/pattern recognition. The most effective approaches of noise reduction are non - linear and adaptive in nature. The ability of the filter unwanted impulse noise and the random noise while preserving edge and details of an image is a non trivial task. Various nonlinear filters based on classical techniques and fuzzy techniques have emerged in the few years for this challenging task. Since filtering a region might destroy an edge and sharpening edges might lead to unnecessary noise. The type of algorithm to be used depends upon the objective to be achieved by the filtering process as well as the particular application. The objective of the thesis is development of fuzzy filter, which is adaptive in natu...