An improved fuzzy rule-based segmentation system

In this paper, we present an improved segmentation process. The presented method is based on the fuzzy rule development obtained from membership functions and applied to roads maps. Features are determined to operate a reliable segmentation. We make use of three features, difference intensity, standard deviation and a measure of the local contrast to classify each pixel to the foreground, which consists of character and line patterns, and to the background. K-means algorithm is used to cluster features vectors. The computed parameters are translated into linguistic variable developing the fuzzy rules system representing segmentation process. Two methods based on k-means algorithm are developed. The first method constitutes a pre-processing for the second method. It permits to select pertinent parameters and adapt their structure for a better detection of information. It doesn´t require a training phase.