摘要
提出了一种新的结合实数编码遗传算法的模糊阈值分割方法。结合遗传算法内在并行运算的特点,此方法在选取多阈值时的效率明显高于传统的模糊阈值法。适应度函数中引入一个新的衡量分割结果的连通性的因子———连通度,克服了传统阈值方法中未考虑像素空间拓扑关系的缺陷。实验证明,此方法比传统模糊阈值方法在运行效率和分割子区域的空间连通性上都有很大程度的改进。
A modified fuzzy thresholding method based on real-coded genetic algorithm is proposed in the paper. With the inherent parallel computing ability of genetic algorithm,the proposed method performs more efficiently in multithreshold selection than traditional fuzzy thresholding. Together with fuzziness,connectivity measure, a new measurement factor,which can be used to evaluate the connectivity of image segmented by the current gray threshold vector,is introduced into the adaptability function. Experimental results show that the proposed method outperforms traditional fuzzy thresholding method in terms of efficiency and spatial connectivitiy.
出处
《计算机应用与软件》
CSCD
北大核心
2006年第11期11-13,共3页
Computer Applications and Software
基金
国家"863"项目(2002AA71610)
关键词
模糊同值法
实数编码的遗传算法
模糊度
连通度
Fuzzy thresholding Real-coded genetic algorithm Fuzziness Connectivity measure
