Identification and generalization of area clusters by mathematical morphology
面状分布地物群识别与概括的数学形态学方法

Many kinds of area features tend to distribute in clusters in map sheets.These clusters usually impress map reader with the concentration of map features.When a generalized map is derived from a detailed map,the pattern and characteristics of cluster distribution should be maintained to some extent.This paper discusses the method of identification and generalization of clusters and regular structural units by mathematical morphology.Clusters are identified based on two criteria,i.e.,proximity between features inside clusters and the ratio between distances inside and outside cluster.Further more,neighboring features with similar size,spacing and orientation are identified as regular structural units inside clusters.After cluster identification,regular structural units are generalized in a way to maintain their characteristic pattern on size,spacing and orientation.Other features inside clusters are generalized associated with theircontext,so thatthe general characteristics and distribution of clusters are maintained.