一种基于紧密蔓延树的形状区域数据聚类算法

摘要

A density spread tree clustering algorithm ( DSTC) was proposed for data from areas of different shapes based on the features of clustering data and making use of the strength of the clustering algorithms available. A series of clustering tests were conducted to the algorithm on data and traffic sign images of areas of different shapes. They demonstrate that DSTC can cluster and classify data from areas of different shapes.%从聚类数据特点出发，分析了现有的各种聚类算法的特点，提出一种基于密度分布的紧密蔓延树聚类（ density spread tree clustering algorithm ， DSTC）算法，该算法由构建紧密蔓延树和数据聚类两部分构成。在各种形状区域数据和交通标志图像数据上对DSTC算法进行了聚类能力验证，实验结果表明， DSTC算法能够聚类区分不同形状区域中的数据。