Robustness Of Order Statistic Filter And Application

Applications of nonlinear filter are springing up in engineering. In some cases we can not get good results by linear filters, but nonlinear filters is fit for that. 11 is difficult to find a common method for constructing the optimal nonlinear filter under some criterion. We should discuss the feature of nonlinear filter according the differences among them.Order statistics play an important role in statistics. Robustness is a criterion for the feature of estimates. Here, we discuss the properties of some nonlinear filters based on the robustness of order statistics. The properties can be applied in document analysis.We divide document image into two parts-text and background. Text area is regarded as a series of random variables with 0-1 distribution and background area as a series of random variables with degenerate distribution. Because the median filter is robust for degenerate distribution, it can reduce the noise. The range is not robust for any bounded random variable, so it can be applied in edge detection of document analysis. Computing results demonstrate the effectiveness of these methods.