Generalized fuzzy c-means with spatial information for clustering of remote sensing images

Fuzzy c-means clustering technique has been popularly used for remote sensing image data classification. However as per the studies the classical fuzzy c-means clustering algorithm has been able to achieve less accuracy due to spatial relationship existence and multi class existence in remotely sensed images. Remote sensing images contain large number of classes but the probability of a pixel belonging to some classes may be low. Traditional fuzzy c-means algorithm considers all classes simultaneously during clustering process. In this paper generalized fuzzy c-means has been applied in exploring k nearest neighbors approach out of c cluster centers. Spatial information has been also integrated with generalized fuzzy c-means technique. The experimental results show that the generalized fuzzy c-means technique with spatial information yields better results than traditional fuzzy c-means technique.