摘要
有效的图像分割是进一步进行图像高层次理解和应用的基础。遗传算法是一种鲁棒性很强的优化算法。该文利用遗传算法对图像进行聚类分析,提出了两种新颖的图像分割算法。染色体码长固定时,按用户指定的特征向量在特征空间内进行聚类分割;染色体码长可变时,可同时对图像应分类数进行动态优化。通过实验对它们各自的优缺点进行了分析,并与其它分割算法的性能进行了比较。利用VisualC++6.0实现了文中的算法,并设计了一具体的小型应用系统。
An efficient color image segmentation method is a foundation to the comprehension and application of image.Genetic Algorithm is a strong robust optimizing method.In this paper two novel genetic clustering color image segmentation algorithms are proposed.In the first algorithm the chromosome's length is fixed.In the second one the length can be optimized with the population's evolution.So the number of regions in color image segmentation can be achieved dynamically from the second algorithm.These algorithms are realized by Visual C++6.0and are compared with other existed results.Finally,an application system is designed.
出处
《计算机工程与应用》
CSCD
北大核心
2003年第27期87-90,192,共5页
Computer Engineering and Applications
基金
国家自然科学基金项目(编号:60174042)
山东省博士基金项目
关键词
遗传算法
聚类分析
彩色图像分割
遗传聚类
可变码长
演化策略
Genetic Algorithms ,Clustering Analysis,Color Image Segmentation,Genetic Clustering,Variable Code Length,Evolution strategy
