中文摘要：

鉴于人工蜂群算法 （ Artificial Bee Colony algorithm, ABC）常用于连续域，具有控制参数少、操作简单和易于实现等优点和二维最小误差阈值分割法复杂度高的问题，提出一种基于人工蜂群算法的二维最小误差阈值分割方法（ Two-dimensional Minimum Error Thresholding based on Artificial Bee Colony algorithm, TMET-ABC）。针对离散域的二维最小误差阈值分割方法，对标准的ABC算法进行修改，构建离散域的人工蜂群算法；依据最小误差公式中变量的三种不同计算方式，提出了3种TMET-ABC方法：直接计算TMET．ABC、数组计算TMET—ABC和递推计算TMET—ABC。仿真结果表明，提出的3种方法能够在达到现有的二维最小误差阈值分割法的分割效果同时，大大加快运行速度，并且3种方法可应用于不同的分割场合。

英文摘要：

In view of the high complexity of the two-dimensional minimum error thresholding meth- od, and simple processing, easy realization and a few control parameters of the artificial bee colony algorithm in a continuous domain, a two-dimensional minimum error thresholding method based on the artificial bee colony optimization algorithm （TMET-ABC） is proposed in this paper. Directed at a two-dimensional minimum error thresholding method in a discrete domain, the ABC algorithm was modified in order to create a discrete ABC algorithm. Then, three TMET-ABC algorithms were pro- posed according to three ways to get variables of two-dimensional minimum error thresholding formu- la： direct computing TMET-ABC （DTMET-ABC）, array computing TMET-ABC （ATMET-ABC） and reeursive computing TMET-ABC （RTMET-ABC）. Simulation results show that the three algo- rithms can greatly improve the running speed while the segmented results are as good as the existing two-dimensional minimum error thresholding method and that they can have different segmentation applications.