摘要
差分演化算法的实现简单有效,但其搜索能力较弱,对此提出一种基于贝塔分布的控制参数动态设置策略以提高差分演化的优化效果,并将其应用于图像分割问题。首先,将图像的直方图按强度分为两类,并按类内方差、类间方差与总方差总结为待优化的目标函数;然后,使用改进的差分演化算法搜索图像分割目标函数的最优解,其中在每轮迭代中使用贝塔分布动态的设置控制参数。仿真实验表明,该方法获得了较好的优化结果,并获得了较好的图像分割效果。
The differential evolution algorithm is effective and easy to realize,but it has poor search ability,so a controlparameter dynamic setting strategy based on beta distribution is proposed to improve the optimization effect of the differential evo?lution,and applied to the image segmentation. In the scheme,the image histograms are divided into two classes according theirintensity,and summarized to the waiting optimization target function according to the inner?class variance,inter?class varianceand total variance. And then,the improved differential evolution algorithm is used to search the optimal solution of the imagesegmentation target function,in which the beta distribution is used to set the control parameters dynamically in each iteration.The simulation experiment results show that the proposed method can obtain better optimal result and good image segmentationeffect.
作者
范泽华
白铁成
FAN Zehua;BAI Tiecheng(College of Information Engineering,Tarim University,Alar 843300,China)
出处
《现代电子技术》
北大核心
2016年第14期87-91,共5页
Modern Electronics Technique
基金
国家自然科学基金资助项目(41561088)
关键词
贝塔分布
差分演化
图像分割
阈值化分割
控制参数
beta distribution
differential evolution
image segmentation
thresholding segmentation
control parameter
