In this paper we propose a new clustering approach for very large spatial datasets that are heterogeneous and distributed. The approach is based on K-means Algorithm but it generates the number of ...