中文摘要：

提出了一种复杂背景下红外目标分割的有效方法。该方法首先利用meanshift的自适应平滑滤波特性，在不损失目标信息的情况下，滤除复杂背景的杂波干扰；然后根据滤波得到的区域，用蚁群方法在区域之间进行聚类合并，获得最终的分割结果。采用区域来表征蚂蚁，与基本蚁群算法将每个像素看作一只蚂蚁相比，其蚂蚁个数大大减少，因而减小了计算的复杂度，提高了图像处理的效率。在蚁群算法中引入了一种新的引导函数，可以更准确引导蚁群聚类。实验结果表明，该方法可以准确地分割出目标，是一种快速有效的图像分割方法。

英文摘要：

A novel algorithm was presented to achieve an improved complex infrared objects segmentation performance. First, the algorithm uses discontinuity preserving smoothing algorithm based on mean-shift procedure to filter the powerful noise without the loss of the object information. Second, the regions produced by mean-shift filtering were merged by ant colony clustering algorithm to gain result of image segmentation. Due to the less ants produced by the regio~ s of filtered image than the original image, and a new visibility based on intensity distribution is defined, it＇s more accurate and efficient to cluster ant colony. Experimental results show the superior performance of the proposed infrared object segmentation algorithm.