基于形态学和蚁群聚类的台风分割方法研究

Research on Typhoon Segmentation Based on Morphology and Ant Colony Clustering

蚁群算法是具有离散性、并行性、鲁棒性、正反馈性等特点的一种仿生进化算法。根据数字图像的离散性特点,从聚类角度出发,将蚁群聚类算法引入到台风云系分割中,并通过引入初始聚类中心和引导函数来解决传统蚁群聚类算法计算量大,搜索时间过长的问题。在蚁群聚类算法的基础上,针对单纯采用蚁群聚类算法可能会将类似于台风云系的分布不均匀的无关云团分割出来而造成分割精度低的情况,提出了一种蚁群算法融合数学形态学方法的台风云系分割方法。仿真结果表明此方法可以准确的分割出台风云系,是一种有效的图像分割方法。

As a bionic evolutionary algorithm,ant colony algorithm has the characteristics of discrete,parallel,robust and positive feedback.According to the discrete characteristics of digital images,the ant colony clustering algorithm was applied to the segmentation of the typhoon from view of the clustering.By introducing the initial clustering centers and guidance function,the large calculation and the search time of traditional ant colony clustering algorithm could be obviously shorten.The problem of low accuracy might be caused by the uneven distribution of unrelated clouds which was similar to typhoon when the ant colony algorithm was utilized respectively.To solve the problem,ant colony algorithm integration of mathematical morphology of typhoon segmentation was proposed.Simulation results show that this method can accurately split introduced typhoon and it is an effective method of image segmentation.