改进人工蜂群算法在图像分割中的应用

作为图像处理技术的一个分支,多阈值图像分割技术已经越来越吸引人们的注意.然而,很多阈值分割技术计算时间较长,且其随着维数的增加而呈指数性增长.因此,为了提高分割的效率,引入基于改进人工蜂群优化算法的多阈值图像分割技术.在分析了标准人工蜂群算法缺陷的基础之上,从雇佣蜂和观察蜂的搜索公式进行改进,使其能够更有效率地收敛至全局最优,同时采用最大类间方差法(Otsu)作为测试改进算法性能好坏的标准.实验证明,改进后的算法更好地平衡了全局搜索和局部寻优能力,在加快收敛速度的同时提高了寻优精度,获得了良好的图像分割效果.

Multilevel image segmentation based on an improved artificial colony algorithm

As a branch of image processing, multilevel image segmentation has attracted more and more attentions. However, most of the threshold image segmentation techniques are time consuming and their computation time grows exponentially with dimension increasing. To improve the efficiency, an improved artificial bee colony algorithm (ABC) for the segmentation problem is introduced. Based on analyzing the shortcoming of the traditional ABC, the searching equation of employed bees and onlooker bees are improved for achieving a precise convergence to the global optima. Experimental results on a set of test images demonstrate that the proposed algorithm makes a better balance between exploration and exploitation, which gets higher solution accuracy and accelerates the convergence speed.