中文摘要：

目的为了解决图像识别中类别划分及划分性能问题,提出基于粗糙集和蚁群优化方法相融合的图像聚类分析方法。方法先对所有样本进行编号并提取数字特征后,采用蚁群优化方法进行数字聚类,最后采用基于粗糙集理论的正域关系进行聚类结果评价。结果与遗传算法实验比较,所提出的方法能获得更理想的聚类结果。结论利用蚁群算法的寻优能力和粗糙集的不确定信息评价的优势,对数字图像进行聚类分析是有效的。

英文摘要：

Aim To solve the classification and performance problems of image recognition,a clustering analysis method is proposed for image clustering by combining the rough set theory and the ant colony algorithm.Methods First of all,the total samples were numbered and corresponding digital features were extracted.Then they were clustered by using the ant colony algorithm.Finally,the performance of the clustering was evaluated by using the principle of the positive domain of rough set theory.Results Compared with the method based on GA,the proposed method can reach better clustering performance.Conclusion The advantages of ant colony algorithm in optimization ability and rough set in dealing with information,are greatly helpful for realizing effective clustering analysis of digital images.