A fuzzy segmentation-based approach to extraction of shorelines from IKONOS imagery. Geomatica 62:407–417

This paper presents a new approach to effective and accurate extraction of coastlines from commercial high-resolution satellite imagery. The approach is based on a fuzzy segmentation algorithm, which models segmentation of a colour image as a fuzzy c-partition and uses colour histogram analysis and a genetic algorithm to determine the number of clusters and optimal fuzzy c-partition, respectively. After segmenting the satellite image into homogenous regions based on colour similarity, the segmented regions are classified into water and land semiautomatically. Then, morphological operations are applied to filter noise in the classified images. The coastline is finally extracted from the filtered images by a delineation algorithm. The developed approach is tested by using 1 m resolution IKONOS pan-sharpened imagery. The effectiveness of the algorithm is demonstrated by the experimental results and accuracy evaluation. 1