模拟退火与模糊C-均值聚类相结合的图像分割算法

Image Segmentation Algorithm Based on Simulated Annealing and Fuzzy C-Means Clustering

模糊C-均值(FCM)聚类算法是一种结合无监督聚类和模糊集合概念的图像分割技术,比较有效,但存在着受初始聚类中心和隶属度矩阵影响,可能收敛到局部极小的缺点。将模拟退火算法(SA)与模糊C-均值聚类算法相结合,在合理选择冷却进度表的基础上,依据模糊C-均值聚类算法建立模拟退火算法的目标函数,实现了基于模拟退火的模糊C-均值聚类图像分割算法。实验表明,该方法具有搜索全局最优解的能力,因而可得到很好的图像分割结果。

The fuzzy c-means （FCM） clustering algorithm is an effective image segmentation algorithm. But it is sensitive to initial clustering center and membership matrix and likely converges into the local minimum, which causes the quality of image segmentation lower. A new image segmentation algorithm is proposed, which combines the simulated annealing （SA） and FCM clustering. The objective function is set up according to FCM clustering and a reasonable cooling schedule is chosen for SA procedure. Some experiment results are given, which show that the algorithm has the effective ability of searching global optimal solution.