| 基于蚁群算法的图像分割方法改进研究 |
| |
| 引用本文： | 张健.基于蚁群算法的图像分割方法改进研究[J].湖州职业技术学院学报,2014(1):84-87,91. |
| |
| 作者姓名： | 张健 |
| |
| 作者单位： | 同济大学软件学院,上海普陀区200331 |
| |
| 摘 要： | 图像分割是一种对不同特征的像素进行聚类的过程,过程中涉及像素的梯度、灰度及邻域特征。由于蚁群算法的离散性、并行性、全局优化性和稳定性等特点,基于蚁群算法提出一种有效的图像分割改进算法。首先通过蚁群改进算法的模糊聚类能力可以分别计算出像素与目标、背景、噪声点、边界点的隶属关系;然后对于蚁群算法循环次数多、计算量较大的问题,设置启发式引导函数和聚类中心,解决传统聚类中运行效率低、初始化敏感度高等缺点;最后引入梯度算子,对目标与背景灰度值相似图像进行分割,对结果进行了进一步的优化。实验表明,应用该改进算法得到的图像分割结果具有较高的准确度和效率。
|
| 关 键 词： | 蚁群算法 图像分割 模糊聚类 梯度算子 |
Improvement Research on Image Segmentation Method Based on Ant Colony Algorithm |
| |
| ZHANG Jian.Improvement Research on Image Segmentation Method Based on Ant Colony Algorithm[J].Journal of Huzhou Vocational and Technological College,2014(1):84-87,91. |
| |
| Authors: | ZHANG Jian |
| |
| Institution: | ZHANG Jian (Software Institute, Tongji University, Putuo 200331, China) |
| |
| Abstract: | Image segmentation is the process of clustering the pixels with different characteristics such as the gradient,grey level and the neighborhood.This paper is to design an effective algorithm in image segmentation using the ant colony algorithm for its characters such as discreteness,parallelism,global optimization and stability.First,the subjection relationship between pixels and target,background,noise as well as boundary membership can be achieved by improved calculating method of ant colony algorithm.Then,aiming to the problem of the more cycle times and a large amount of calculation,this research sets the heuristic function and the cluster centers,so as to solve the tradi-tional clustering of low efficiency,high sensitivity for initialization.Finally,in order to solve the problem of separating the object from back-ground,which has similar gray value with the object,the gratitude operator is introduced into this algorithm.The experiments show that the effect of improved image segmentation is more accurate and effective. |
| |
| Keywords: | ant colony algorithm image segmentation fuzzy clustering gradient operator |
| 本文献已被 CNKI 维普 等数据库收录！ |
|
