期刊名称:ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN:2194-9042
电子版ISSN:2194-9050
出版年度:2008
卷号:XXXVII Part B7
页码:565-568
出版社:Copernicus Publications
摘要:For a clustering algorithm, the number of clusters is a key parameter since it is directly related to the number of homogenous regions in the given image. Although ISODATA clustering algorithm can determine the number of clusters and cluster centers dynamically, it is challenging to specify so many parameters. The possibility clustering provides the memberships that are interpreted as the degrees of possibility. So the memberships in one class are not related to the memberships in other classes. By taking the advantages of these two clustering algorithms, a new fuzzy clustering algorithm is proposed by combining the possibility clustering and ISODATA clustering algorithm. This new algorithm not only can determine the number of clusters dynamically with the degree of possibility of each date point, but also can reduce the number of input parameters of ISODATA algorithm. The splitting and merging process is evaluated by the possibility distance