二维最大熵法和二维最小交叉熵法是目前常用的两种阈值分割方法, 但在某些时候因为两种方法获取的阈

值过高或者过低, 使得分割失效。针对此问题, 提出了基于二维最大熵法和二维最小交叉熵法结合的图像分割方法。

首先, 对二维最小交叉熵公式进行转化 然后, 利用多目标规划理论将这两种方法有机结合使得到的阈值既满足二维

最大熵原则, 又满足二维最小交叉熵原则 最后, 利用二维直方图的特点推导出新型递推算法搜索最佳阈值并降低计

算复杂度。-The thresho ld ing method based 2-D max imum entropy and the one based on 2-D m inimum cross

entropy are used w ide ly in im age segmentation today, bu t in some app lications, they fa il to segment im ages

because of too h igh or too low thresho lds. Therefore, we proposed an im age thresho ld ing m ethod based on

the comb ination of 2-D max imum entropy and 2-D m inimum cross entropy. Firstly, the formula of the 2-D

m in imum cross entropy w as transfo rmed, then 2-D max imum entropy and 2-D m inimum cross entropy w ere

combined together usingmult-i objective prog ramm ing theory so that the optima l thresho ld value could satisfy

the threshold requirem ent of the bo th. A new recursive algo rithm w as in ferred using the features o f the 2-D

h istogram in order to search the best thresho ld vecto r and to reduce the compu ting complex ity