一种基于多重模糊聚类的红外目标分割算法

提出了一种基于多重模糊聚类的红外目标分割算法。为了实现目标的准确分割,先将原始红外图像进行四划分得到四个子图像,在各个子图像上分别进行模糊C均值聚类,再对图像进行横纵二划分各得到两个子图像,并将四划分时得到的聚类结果约束在二划分的聚类过程中,最后将二划分得到的聚类结果约束到原始图像的聚类过程中,并在其中加入邻域空间约束。此方法可有效增强背景和目标区域像素点的各自凝聚性和抗干扰性,有效提高聚类分割结果的准确性。实验结果表明,多重模糊聚类目标分割算法能准确地实现红外图像目标区域和背景区域的分离,是一种可行的目标分割算法。

Algorithm for infrared target segmentation based on multiple fuzzy clustering

An algorithm for infrared target segmentation based on multiple fuzzy clustering is proposed.In order to segment target accurately,the original image is divided into four parts.The clustering is performed on these four parts.Then the original image is divided into two parts.The clustering procedure is constrained by the results in former clustering results.Finally,the clustering procedure of the whole image is also constrained by the results in former clustering results.And the space restriction of neighborhood is added.This method can enhance the coagulability and anti-interference performance of background and target area,which make the target segmentation accurate.The experimental result shows that the target area and background area can be separated well in infrared image by the proposed method,which is a feasible method for target segmentation.