一种基于量子遗传算法的红外图像分割方法

A method of infrared image segmentation based on quantum genetic algorithm

针对经典的二维O tsu和最大熵算法的不足,提出了一种基于量子遗传算法的海上红外图像目标分割方法.该方法同时考虑了图像像素点的灰度分布信息和像素间的空间相互信息,将基于fisher准则的类内类间距离判据作为分类依据,利用量子遗传算法进行寻优以获取最佳阈值,实现了海上红外目标图像的分割过程.选取3幅海上红外目标图像进行了仿真实验.实验结果表明,提出的方法在分割效果和计算速度上都优于传统的O tsu和最大熵法.

Aming at the disadvantages of classical Otsu and maximum entropy algorithm, a method of maritime infrared target segmentation based on quantum genetic algorithm was presented. In this method, both gray distribution information of pixels and inter-space information between pixels were used, the criterion of withinclass and between-class distance was taken as the rule, and quantum genetic algorithm was utilized to search the optimal threshold that realized the classification of target and background. Three maritime infrared target images were selected in the simulation experiments. The experimental results show that the proposed method greatly outperforms Otsu and maximum entropy algorithm in the performances of both computation accuracy and speed.