基于人工免疫的灰度图像多阈值自动分割

Automatic multi-threshold image segmentation based on artificial immune algorithm

为了实现灰度图像的自动分类以及自动分割,提出了一种基于人工免疫及最优分类数的灰度图像多阈值自动分割方法。定义了灰度图像最优分类数目标函数;接着运用人工免疫算法,结合最优分类数函数对灰度图像进行自动分类,并产生最优的多阈值,从而使得图像的全自动分割成为可能。该人工免疫算法中,抗原是指最优分类数目标函数,而抗体是指最优的多阈值。通过实验证明,分类清晰,效果良好。

For the number of clusters is determined automatically so that a multi-threshold image is segmented automatically, an automatic image segmentation method based on AIA （artificial immune algorithm） and optimal number of clustering is presented. First, the objective function for the optimal number of clustering is presented. Then, an artificial immune approach is presented to generate automatically the number of clusters and segmentation thresholds. In this approach, the objective function is regarded as antigens, and the number of clusters and segmentation thresholds are regarded as antibodies. Experiments demonstrate the good performance of the proposed method.