中文摘要：

提出了一种基于区域生长和蚁群聚类的图像分割方法——BRGAC。该方法首先用区域生长法对图像作初始分割,然后利用蚁群算法搜索最优解的能力,在区域之间进行聚类合并,获得最终的分割结果。BRGAC算法不但克服了区域生长得不到有意义区域的不足,而且还大大提高了蚁群聚类算法的搜索时间,并利用初始分割后的空间信息和灰度信息定义了一种新的引导函数,可更准确有效引导蚁群聚类。实验结果表明,该方法可以准确地分割出目标,是一种有效的图像分割方法。

英文摘要：

This paper proposed an image segmentation method based on region growing and ant colony clustering. First, the image was segmented with region growing algorithm ; considering the good performance of BRGAC in searching the best solution of ant colony, it was used to merge different regions of homogeneity and gain the result of segmentation. BRGAC algorithm could overcome the disadvantage of region growing which can＇ t get meaningful region. Besides, it could reduce the computational time of ant colony clustering. What＇s more, BRGAC defines a new visibility based on intensity and spatial information from initialization segmentation, which guides ant clustering more effectively. Experimental results show that BRGAC can segment the image accurately and precisely, and thus is a effective method for the image segmentation.