一种基于粒子群优化的K-均值彩色图像量化算法  

A K-mean color image quantization method based on particle swarm optimization

作　　者：许永峰[1,2] 姜振益[1] 

机构地区：[1]西北大学现代物理研究所 [2]西北大学数学系,陕西西安710069

出　　处：《西北大学学报（自然科学版）》2012年第3期351-354,共4页Journal of Northwest University（Natural Science Edition）

基　　金：国家自然科学基金资助项目(10647008;50971099);教育部博士点基金资助项目(20096101110017);陕西自然科学基金资助项目(2010JZ002);陕西省教育厅科研基金资助项目(09JK782)

摘　　要：目的利用粒子群优化算法和K-均值方法研究彩色图像的量化问题。方法针对K-均值聚类量化算法对初始值比较敏感,易陷入局部极小值从而使得算法得不到全局最优解,为局部搜索算法,以及粒子群优化算法是一种全局寻优方法的特征,把K-均值聚类方法和粒子群优化算法结合起来,将K-均值聚类方法中的聚类函数作为粒子群优化算法中的粒子适应度函数,对彩色图像进行聚类量化。结果实验表明新算法在峰值信噪比和均方根误差评判准则下可以得到更好的量化结果。结论新方法有效地克服了K-均值聚类方法和粒子群优化算法的不足。Aim To adopt PSO and K-mean to study the process of color image quantization. Methods K-mean clustering algorithm is a local search algorithm because it is easily trapped local optimum and is sensitive to initial value effectively. On the other hand, particle swarm optimization algorithm （PS0） is a global optimization algo- rithm. By incorporating the local search ability of K-mean algorithm and the global optimization ability of PSO and taking the criterion function of K-mean as the object function of PSO, a new hybrid color quantization algorithm based on PSO and K-mean algorithm is proposed. Results It was proved by the experiments that the new algorithm can get the optimal quantized image by PSNR and RMSE. Conclusion The new method effectively overcomes the insufficiency of K-mean and PSO.

关 键 词：彩色图像量化 K-均值聚类 粒子群优化 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]