摘要
彩色图像数据信息量较大,传统的模糊C均值聚类算法(FCM)在分割时更加容易受到初始聚类中心影响陷入局部极值.文中研究了一种融合差分演化、粒子群和模糊均值聚类的彩色图像分割算法(DEPSO-FCM).利用差分演化算法的快速收敛特性、粒子群算法的全局搜索能力,解决模糊均值聚类图像分割时易受到初始聚类中心影响和陷入局部最优的问题,同时针对不同的色彩空间对于图像分割效果的影响,尝试在不同的空间上使用DEPSO-FCM进行图像分割.实验表明,该方法能解决FCM算法陷入局部最优的问题,在不同的色彩空间上都获得了理想的分割效果.
Color image contains a lot of data, the traditional fuzzy C-means clustering algorithm (FCM) easy to fall into local extremum by the influence of the initial cluster centers in segmentation. This paper presents a hybrid differential evolution (DE), Particle swarm optimization (PSO) and fuzzy C-means clustering (FCM) algorithm called DEPSO-FCM for color image segmentation. By the use of the fast convergence of DE and the global search ability of PSO to solve FCM color image segmentation which is influenced by the initial cluster centers and easily into a local optimum. Meanwhile, considering the influence of different color space, using DEPSO-FCM split images in different color space. Experimental results show that this method can solve the problem of local optimum, and can get ideal color segmentation results in the different color st^ace.
出处
《江西理工大学学报》
CAS
2013年第5期66-71,共6页
Journal of Jiangxi University of Science and Technology
基金
江西省教育厅青年科学基金项目(GJJ13377)
关键词
模糊C均值聚类
差分粒子群算法
全局优化
彩色图像分割
fuzzy C-means clustering
differential evolution particle swarm optimization
global optimization
color image segmentation