Comparision of Clustering Methods in 2D Image for the Atomstion of Dangerous Machine

위험기계의 자동화를 위한 2차원 영상의 군집화 기법 비교 연구

In this study, clustering of black BADUK stones' image were performed to recognize the individual stone from its closely spaced and partially occluded Image. And the clustering perfomance was compared between the classical methods and fuzzy C-means method. As a result, 2 BADUK stones' image was segmented precisely in every methods, but more than 3 stones the segmentation was depended on its shape. Fuzzy C-means method could be segmented correctly to 4 stones regardless of its shape, and It could be applied to the unknown number of clusters.