基于参数自适应粒子群模糊聚类的图像分割研究

Research on image segmentation based on parameter adaptive particle swarm fuzzy clustering

传统模糊聚类算法初始值过多依赖先验知识,易陷入局部最优,而采用粒子群模糊聚类虽有所改善,但粒子群也存在陷入早熟收敛的情况。针对这一问题,提出一种基于参数自适应的粒子群模糊聚类算法APSO-FCM。首先采用自适应参数控制策略对粒子群算法进行改进,然后使用提出算法优化模糊聚类初始数目,接着使用新的模糊C均值聚类方法进行图像分割。实验结果表明,提出算法划分的图像效果得到改善,在划分系数、划分熵等图像分割指标上有进一步提升。

The initial value of traditional fuzzy clustering algorithm relies too much on prior knowledge,which is easy to fall into the local optimum.Although the use of particle swarm fuzzy clustering has been improved,the particle swarm also has the situation of falling into premature convergence.Regarding this problem,a particle swarm fuzzy clustering algorithm based on parameter adaptation,APSO-FCM,is proposed.To begin with,the algorithm uses an adaptive parameter control strategy to improve the particle swarm algorithm,then the new algorithm is used to optimize the initial number of fuzzy clusters,and the new fuzzy C-means clustering method is adapted for image segmentation.The experiment results show that the image effect of the new algorithm is improved,and the image segmentation indexes such as the partition coefficient and the partition entropy are further improved.