基于进化规划算法的图像聚类研究

Image clustering based on evolutionary programming algorithm

将进化规划算法应用于图像聚类问题,对问题的解进行符号编码,采用群体智能模式实现问题解的搜索.利用进化规划算法的变异算子和选择算子可以有效提高算法的全局搜索能力,采用高斯变异算子保证了优秀解的多样性,降低了进化操作的复杂性.仿真实验证明基于进化规划算法的图像聚类算法具有可行性和准确性.

The evolutionary programming algorithm was applied to the image clustering problem, the solution of the problem was encoded by the symbol coding, the search for the solution of the problems was achieved using swarm intelligence mode. The global search ability of the algorithm was improved by mutation operator and selection operator of the evolutionary programming algorithm. The diversity of the excellent solution was ensured and the complexity of evolutionary operation was reduced by Gauss mutation operator. The feasibility and accuracy of the image cluster algorithm based on evolutionary programming algorithm were proved by the simulation experiments.