An integrating KFCM and threshold segmentation fusion method for contour lines extraction from color scanned topographical maps

Extracting contour lines from digital maps is a difficult problem which is often encountered when building a geographic information system. In this paper, a fusion algorithm which integrates kernel fuzzy C means (KFCM) clustering technique and threshold segmentation technique is presented to automatically extract the contour lines. Firstly, color image of scanned topographical map is converted into HIS color space. Then its corresponding feature image is constructed by using its saturation channel data from HIS space and its space information image together. Finally segmentation threshold value is calculated by applying KFCM clustering algorithm and threshold segmentation method is implemented to extract contour lines. Experimental results show that the fusion algorithm can overcome the segmentation difficulties from the color diffusion phenomenon. No matter the distribution of grayscale image is unimodal, bimodal, or multimodal, the method can effectively extract the contour lines from color map images.