Study On Image Processing Based On Rough Sets Theory

Rough sets theory, proposed by Professor Z.Pawlak in 1980s, is a theoretical method that researches not only vagueness and uncertainty of knowledge, but also expression and induction of data. Before and after 1990, rough sets theory has gradually attracted worldwide attention because of its successful applications in data decision-making and analysis, pattern recognition, data mining, machine learning and knowledge discovering. At present, image processing methods based on rough sets theory have become the new investigative direction of image processing. The main work of this paper is to explore some applications of rough sets theory that includes image clustering, image edge-detecting and mathematical morphological. So far, though studies on this field have existed, but there's a lack of the reference literature and the background of application. This paper wants to do some work on it: one is to explore the potential of rough sets theory in application; the other is to find a new image processing method.Attribute reduction is the core of rough sets theory. In this paper, reduction algorithm based on heuristic information, which is efficient and simple, is presented firstly. The algorithm is proved to have good performance by experiments on UCI datum. Based on rough sets theory, classical fuzzy C-means algorithm is ameliorated. A new rough fuzzy C-means algorithm (RFCM) is proposed and the better effect can be testified by experiments both on the performance of clustering and the initial value sensitivity. Furthermore, a fuzzy rough set model's applied in image processing is investigated. A new binary fuzzy rough sets model based on triangle modulus is discussed, and its application in image edge-detecting and noise eliminating shows better affect than classical edge-detecting operators, such as sobel, prewitt, Rorbets and so on. In addition, compared with the classical mathematical morphology, a new mathematical morphologic algorithm based on rough sets not only ensures continuity of the edge, but also renders the edge thin. Finally, the application of rough sets in image processing is summarized and the direction of further research is expected.